Prematurity adds a second developmental risk signal after infant heart surgery · NICU360
Research
Prematurity adds a second developmental risk signal after infant heart surgery
NICU360 Evidence Desk·7 min read·00
Among 513 children who had infant cardiac surgery and returned for Bayley-III testing, earlier gestation tracked lower adjusted cognitive and motor scores. Only 38% of 1,350 eligible children returned, so the signal is clinically important but not an individual forecast.
THE HOT TAKE
Gestational age is a follow-up signal here, not a causal lever and not an individual forecast.
A 2026 single-center retrospective cohort linked prematurity with lower toddler cognitive and motor scores after infant cardiac surgery for congenital heart disease. Of 1,350 eligible children, 513 returned for Bayley-III assessment. In the final models, children born at 33 weeks or earlier scored 11.2 cognitive points and 15.2 motor points lower, on average, than the group born at 39 weeks or later. The abstract summarizes the corresponding differences as about 11 points (95% CI 5-17 lower) and 15 points (95% CI 8-22 lower).
Adjusted Bayley-III score deficit versus birth at 39 weeks or later (points lower)
Cognitive
Language
Motor
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.
33 weeks or earlier (n=21)34-36 weeks (n=44)37-38 weeks (n=136)
These bars convert negative regression coefficients into positive magnitudes of points lower than the 39-week-or-later reference group; they are not raw means, risks, or individual predictions. After adjustment, the motor coefficients met P<0.05 in all three earlier-gestation groups. None of the language coefficients met that threshold, and the cognitive coefficient did so only in the 33-weeks-or-earlier group.
Gestational-age groups and final adjusted coefficients
Gestational age
Children
Cognitive vs reference
Language vs reference
Motor vs reference
33 weeks or earlier
21
-11.2 (P<0.001)
-5.8 (P=0.15)
-15.2 (P<0.001)
34-36 weeks
44
-3.5 (P=0.14)
-3.6 (P=0.25)
-6.6 (P=0.011)
37-38 weeks
136
-2.8 (P=0.052)
-3.1 (P=0.088)
-4.8 (P=0.002)
39 weeks or later
312
Reference
Reference
Reference
WHAT WAS ACTUALLY STUDIED
The cohort came from Boston Children's Hospital. Eligible children were born from April 2007 through March 2020, underwent cardiac surgery in infancy with or without cardiopulmonary bypass, and later completed outpatient Bayley-III testing. Children with trisomy 21 and those whose only operation was patent ductus arteriosus ligation were excluded. This was an association study, not a trial of timing, surgery, developmental care, or follow-up.
From surgical cohort to adjusted association
1
1,350 eligible surgical infants
Congenital heart disease, infant cardiac surgery, birth years 2007-2020.
2
513 returned for Bayley-III
Thirty-eight percent of the eligible cohort completed outpatient developmental assessment.
3
Four gestational-age groups
21 at 33 weeks or earlier; 44 at 34-36; 136 at 37-38; 312 at 39 weeks or later.
4
Latest assessment entered the models
Cognitive, language, and motor scores were modeled with selected demographic, medical, surgical, and postoperative covariates.
Testing occurred from 6 to 40 months of age, and a licensed clinical psychologist administered the Bayley-III. If a child had more than one assessment, the latest was analyzed; 80% were tested after 24 months. Scores for children born before 37 weeks were corrected for gestational age until age 2 years. Small amounts of missing outcome data were excluded from the corresponding analyses.
THE 62% WHO DID NOT RETURN ARE THE MAIN GUARDRAIL
Only 513 of 1,350 eligible children returned. Returnees were more likely to be White and non-Hispanic, more likely to have an extracardiac anomaly, and less likely to have single-ventricle anatomy. The smallest gestational-age group contained just 21 children, none with single-ventricle anatomy. Those selection patterns can shift both the measured scores and the apparent size of the gestational-age association.
The final models explained 22.9% of cognitive-score variance, 17.0% of language variance, and 25.7% of motor variance. In the final hierarchical models, gestational age accounted for 2.6%, 0.8%, and 4.5% of variance, respectively. Caregiver education, race, genetic diagnosis, cardiac anatomy, and postoperative length of stay also appeared in at least one model. These variables are entangled with biology, illness severity, care, and structural conditions; regression does not separate them into causes.
The study also cannot speak for infants who never underwent surgery because they died, received palliation, or were not surgical candidates. Brain MRI was not routine, residual confounding and collinearity remain, and Bayley-III scores in toddlerhood are developmental measurements rather than fixed long-term diagnoses. Nothing here establishes that changing the timing of cardiac surgery would change neurodevelopment.
HOW THIS FITS THE EVIDENCE
A 2024 systematic review found 19 observational studies of preterm or growth-restricted children with congenital heart disease, only six suitable for meta-analysis; eight of 19 were judged at low risk of bias. Its conclusion was appropriately cautious: the evidence that prematurity or growth restriction accentuates specific neurodevelopmental outcomes was low level. The new cohort adds a larger adjusted analysis, but its 21 children born at 33 weeks or earlier do not close that evidence gap.
The 2024 American Heart Association scientific statement already frames congenital heart disease neurodevelopment through combined genetic, fetal, perinatal, surgical, socioeconomic, and family risks, with age-based evaluation for those at high risk. This paper supports gestational age as one useful risk signal within that stack. It does not justify reducing a child's trajectory to gestational age or a single Bayley score.
The practical systems signal is follow-up access. In a separate 16-center cohort of 2,385 eligible children after infant cardiopulmonary bypass, only 692 (29.0%) attended a toddler neurodevelopmental evaluation in the target window. Hospital-initiated rather than family-initiated scheduling was the strongest attendance predictor (adjusted OR 4.24, 95% CI 2.74-6.55). That association does not prove a scheduling intervention works, but it makes loss to follow-up an operational outcome worth measuring alongside developmental scores.
For neonatal and cardiac teams, the proportionate interpretation is to keep prematurity visible when building multidisciplinary follow-up pathways and counseling about uncertainty. The research signal is strongest for motor outcomes and for the most premature subgroup, but it is a group-level association from a selected survivor cohort. Replication needs prospective multicenter enrollment, active retention, prespecified modeling, brain-imaging context, and longer follow-up.
