Treated late-preterm hypoglycemia looked reassuring at 2 years, not definitive
NICU360 Evidence Desk·7 min read·00
In a 128-infant prospective cohort, adjusted Bayley-4 outcomes at 18–24 months did not differ after transient hypoglycemia was identified and managed using AAP thresholds. Selection, attrition, small severe subgroups, and a short follow-up horizon keep this from proving long-term safety.
THE HOT TAKE
For selected 35- to 36-week infants whose low glucose was detected and treated in the first day, the 2-year signal is reassuring. It is not evidence that severe, persistent, or later-emerging effects are absent.
A multicenter prospective cohort followed infants born at 35 0/7–36 6/7 weeks and weighing at least 2000 g from four New York hospitals. Of the 128 children assessed at a median corrected age of 22.3 months, 81 had at least one low glucose value and 47 remained euglycemic during risk-based screening in the first 24 hours.
Bayley-4 mean scores by neonatal glucose status
Cognitive
Language
Motor
Social-emotional
Was this worth carrying into your next shift?
Your reaction stays connected to your NICU360 profile.
Comments
0
Join the conversation
Sign in to reply, vote, and keep your participation with your NICU360 profile.
Glucose was measured with point-of-care meters. Hypoglycemia meant <40 mg/dL in the first 4 hours or <45 mg/dL at 4–24 hours; severe meant <36 mg/dL, and recurrent meant at least three episodes. Management could include supplemental feeding, buccal dextrose gel, or intravenous dextrose.
The prespecified primary outcomes were continuous Bayley-4 cognitive, language, motor, social-emotional, and adaptive scores. Every unadjusted 95% confidence interval included zero; models adjusted for maternal diabetes, antenatal steroid exposure, and family history of developmental delay also remained nonsignificant.
FOLLOW-UP IS THE FIRST FILTER
The cohort narrowed from 939 screened infants to 661 eligible, 235 consented, and 128 assessed. Reported attrition checks found no differences in gestational age, sex, or glycemic status between completers and noncompleters, but that degree of narrowing still leaves room for selection beyond measured characteristics.
The groups were not exchangeable at baseline. Maternal diabetes was more common in the euglycemic group (42.6% vs 24.7%), whereas antenatal steroid exposure (59.3% vs 40.4%) and NICU admission (22.2% vs 4.3%) were more common in the hypoglycemic group. Family history of developmental delay was less common in the hypoglycemic group (17.3% vs 36.2%). Adjustment addressed three variables, but individual socioeconomic measures were unavailable and residual confounding remains possible.
THE EXPLORATORY CATCH
Severe and recurrent subgroup analyses were underpowered. One additional three-tier analysis found borderline language scores in 43.2% versus 20.9% after severe versus milder or no hypoglycemia (p=0.034); another paradoxically favored the hypoglycemia group for motor scores below 85 (8.6% vs 23.4%, p=0.045). With small subgroups and multiple exploratory contrasts, neither result should carry the main conclusion.
WHY TWO YEARS IS NOT THE ENDPOINT
Context tempers both alarm and reassurance. In CHYLD, 404 assessed children born at 35 weeks or later were treated to maintain glucose at or above 47 mg/dL. At 2 years, hypoglycemia was not associated with neurosensory impairment (RR 0.95, 95% CI 0.75–1.20) or processing difficulty (RR 0.92, 95% CI 0.56–1.51).
At 4.5 years, however, the CHYLD cohort still showed no increase in combined neurosensory impairment (RR 0.96, 95% CI 0.77–1.21) but did show associations with low executive function (RR 2.32, 95% CI 1.17–4.59) and visual-motor function (RR 3.67, 95% CI 1.15–11.69). Those observational associations do not prove causation, but they show why a normal toddler composite cannot close the question.
An exploratory hPOD cohort analysis also resisted a simple narrative. Among 1,197 assessed children, hypoglycemia was associated with neurosensory impairment at 2 years (23% vs 18%; adjusted RR 1.28, 95% CI 1.01–1.60), especially after severe episodes (adjusted RR 1.68, 95% CI 1.20–2.36), while adjusted cognitive and motor mean differences were small at −1.48 and −2.05 points. The association does not establish causation.
How far the finding travels
1
Population boundary
Nursery infants born at 35 0/7–36 6/7 weeks and weighing at least 2000 g; 34-week and medically unstable infants were excluded.
2
Treatment context
Risk-based screening and prompt management under AAP operational thresholds; this was not untreated exposure.
3
Time horizon
Bayley-4 at 18–24 months; higher-order executive and visual-motor functions may emerge later.
4
Selection boundary
Non-English-speaking caregivers were excluded, and 128 of 661 eligible infants completed assessment.
PRACTICAL TAKEAWAY
For clinicians, the useful reading is narrow: these data support measured reassurance after promptly identified, transient hypoglycemia in a selected late-preterm nursery population. They do not justify relaxing surveillance, treatment thresholds, follow-up, or concern for persistent, symptomatic, severe, or recurrent episodes. The key research need is larger, more inclusive follow-up into school age.
Treated late-preterm hypoglycemia looked reassuring at 2 years, not definitive · NICU360
