527 - Uncovering Resilience: Identifying Positive Factors in Extremely Preterm Infants Who Exceed Neurodevelopmental Predictions

Publication Number: 527.6645

Andrea F. Duncan, CHOP/Upenn, Philadelphia PA, PA, United States; Carla Bann, RTI International, Apex, NC, United States; Jane E. Brumbaugh, Mayo Clinic, Rochester, MN, United States; Barbara Chaiyachati, Children's Hospital of Philadelphia, University of Pennsylvania, Philadelphia, PA, United States; Sara B.. DeMauro, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, United States; Susan R. Hintz, Stanford University School of Medicine, Palo Alto, CA, United States; Mary Lauren Neel, Emory University School of Medicine, Atlanta, GA, United States; Kathryn Gustafson, Duke University School of Medicine, Durham, NC, United States; H. Gerry Taylor, Nationwide Children’s Hospital, Columbus, OH, United States; Betty R. Vohr, Women & Infants Hospital/ Brown Medical School, Providence, RI, United States; Nathalie L. Maitre, Emory University School of Medicine and Children's Healthcare of Atlanta, Atlanta, GA, United States; NICHD Neonatal Research Network, National Institute of Child Health and Human Development, Bethesda, MD, United States

Background: Children born extremely preterm (EP, < 28wks GA) are at high risk for neurodevelopmental deficits, compounded by biomedical risks and adverse social drivers of health. Yet, some EP children exceed outcome expectations, suggesting potential factors that support resilience in neurodevelopment.

Objective: To shift the study of neurodevelopment in EP children to a success-based framework by identifying factors after discharge linked to surpassing 2-year scores on Bayley Scales of Infant and Toddler Development (BSITD) otherwise predicted by perinatal variables.

Design/Methods: Analyses included inborn EP infants at NICHD Neonatal Research Network centers (2013-2019) followed at 22-26m corrected age. The primary outcome was surpassing predicted BSITD composite scores (cognitive, motor, language), defined as having a score >½ standard deviation above model prediction. A split-half cross-validation approach was used. Predictive models for BSITD scores were developed in half the children based on biomedical and psychosocial variables (Table 1). Models were then applied to the second half to calculate predicted scores. Children had the primary outcome when their actual scores were >½ SD above predicted scores. Logistic regressions examined relationships between modifiable protective factors after discharge and surpassing predictions.

Results: 4356 children were included. Biomedical predictors of lower BSITD scores included male sex, late-onset sepsis, BPD grade 3, NEC, IVH grade ≥ 3, PVL, ROP stage ≥ 3 or plus disease, and shunt for post-hemorrhagic hydrocephalus (Table 2). Lower scores were associated with psychosocial variables of public/no insurance, minoritized maternal race/ethnicity, single-parent household, and social work involvement. In the validation sample, 590 (27%) exceeded expectations in cognitive, 671 (31%) in language, and 622 (29%) in motor domains. Home nursing and second language exposure were associated with lower odds of surpassing expectations on the language composite. NICU follow-up clinic attendance was the only post-discharge factor associated with higher odds of surpassing expectations on cognitive scores (Table 3).

Conclusion(s): More than 24% of EP children surpassed neurodevelopmental expectations. Use of multiple post-discharge services may reflect higher clinical and social risk profiles, offering opportunities for tailored support. However, NICU follow-up attendance is linked to enhanced cognitive development, supporting the clinical value of these programs beyond research. Prioritizing success factors over deficits may better inform strategies to optimize EP outcomes.

Table 1. Biomedical, Psychosocial, and Potential Modifiable Protective Factors for the total Sample
PAS Abstract Table 1.pdf

Table 2. Regression Modeling of BSITD Composite Scores in the Development Sample using Biomedical and Psychosocial Variables: Development Sample (N=2,178)
PAS Abstract Table 2.pdf

Table 3. Odds Ratios of Having BSITD Composite Scores that Exceeded Predictions using Potentially Modifiable Protective Factors - Validation Sample (N=2,178)
PAS Abstract Table 3.pdf