616 - Weight Trajectories Drive Better Head Circumference Growth in Preterm Infants ≤32 Weeks: Evidence from a Multicenter Latin American Cohort, 2015–2024

Publication Number: 1593.616

Angela B. Hoyos, El Bosque University, Bogota, Distrito Capital de Bogota, Colombia; Pablo Vasquez-Hoyos, Universidad Nacional de Colombia, Bogota, Distrito Capital de Bogota, Colombia; Carlos A. Fajardo, University Of Calgary, Calgary, AB, Canada; Horacio Osiovich, University of British Columbia Faculty of Medicine, Vancouver, BC, Canada; Martha L. Baez-Camacho, “El Bosque” University - “Country “ clinic, Colombia, Distrito Capital de Bogota, Colombia; Carolina Villegas, Hospital regional de alta especialidad Dr. Ignacio Morones Prieto, San Luis Potosi, San Luis Potosi, Mexico

Background: Optimal postnatal growth trajectories for preterm infants remain controversial. Both under/overgrowth are concerning, but head circumference (HC) growth is especially relevant given its correlation with brain mass and neurodevelopment.

Objective: To determine the association between postnatal weight trajectory and HC growth, and to identify thresholds beyond which further weight recovery does not improve HC

Design/Methods: This was a a secondary analysis in a multicenter prospective colected cohort study from EpicLatino, a quality improvement database of 36 NICUs across Latin America and the Caribbean, modeled after the Canadian Neonatal Network. Participating units follow all infants until discharge. We included neonates ≤32 weeks' gestational age admitted at less than 48 hours of birth 2015-2024 and discharged alive. Infants with mayor malformations, missing or implausible weight or HC measurements were excluded.
Weight and HC Z-scores at birth and discharge were calculated using the 2013 Fenton charts. The exposure was the change in weight Z-score from birth to discharge (weight ∆-Z-score), and the primary outcome was the change in HC Z-score birth to discharge diference (HC ∆-Z-score). The association between weight ∆-Z-score and HC growth was assessed using multivariable regression models, adjusting for potential confounders selected on the basis of biological plausibility and significance in bivariate analyses. Adjusted probabilities were estimated and summarized by categories of weight ∆-Z-score.

Results: We analyzed 3,358 infants ≤32 weeks' gestation (see tables and figure). After adjustment for gestational age, NEC, SNAP-II, sepsis (+ blood and/or CSF), length of stay, treated patent ductus arteriosus, 5´ Apgar score, SGA ≤10th percentile, and outborn status, HC growth from birth to discharge remained strongly associated with postnatal weight trajectory. Infants with greater weight ∆-Z-score ≤−2.0 had the poorest head growth, with predicted HC ∆-Z-scores around −1.5 to −1.7. Conversely, those with milder weight decline (−0.4 to 0) showed significantly less impairment (predicted HC ∆-Z-score ≈−0.3). Improvement plateaued beyond this point, as infants with weight recovery above birth Z-scores (>0) did not demonstrate additional head growth benefits. All weight ∆-Z-score groups were statistically significant compared with the reference, except the >0 group.

Conclusion(s): Postnatal weight trajectory is closely associated with HC growth in preterm infants ≤32 weeks. Our findings suggest that achieving weight Z-scores as close as possible to ideal growth is critical to support optimal brain growth.

Characteristics
Sepsis: positive blood or/and CSF, GA: gestational age at birth, AC: antenatal corticosteroids, Chorio: chorioamnionitis, SGA ≤ 10th percentile: small for gestational age at birth, NEC: necrotising enterocolitis, CSF: cerebrospinal fluid, RDS: respiratory distress syndrome, TPN: total parenteral nutrition, IQR: interquartile rage.

Linear prediction of HC ∆-Z-score using weight ∆-Z-score groups
HC: head circumference, ∆-Z-score: Delta Z score between birth and discharge, CI: confidence interval, Std: standard, t: Coefficient/Standard Error, p: statistical associated value (significant if <0.05). The 8 groups of Weight ∆-Z-score define in the text is shown in column 1, the predicted HC ∆-Z-score Birth to Discharge is in column 3. All groups were statistically significant except the group >0.

Predictive HC ∆-Z-score Birth to Discharge from Weight ∆-Z-score Groups
Predicted values were obtained from multivariable regression models adjusted for gestational age, necrotizing enterocolitis, SNAP-II, sepsis, length of stay, treated PDA, 5-minute Apgar, and outborn status. Weight ∆-Z-score was divided into eight categories (from ≤−2.4 to >0). Each point represents the adjusted marginal prediction of HC ∆-Z-score for that group, with 95% confidence intervals. The figure shows that HC growth declined progressively as weight loss increased, improved as trajectories approached zero, and plateaued once weight exceeded the birth Z-score. HC: head circumference, ∆-Z-score: Delta Z score between birth and discharge, CI: confidence interval.