306 - Dietary intake patterns in children born preterm and associated growth and developmental milestone assessments

Publication Number: 3295.306

Catherine J. Fergie, Northwestern/ Lurie, Chicago, IL, United States; Colette M. Gordon, Ann & Robert H. Lurie Children's Hospital of Chicago, Park Ridge, IL, United States; Suna Paek, Ann & Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL, United States; Preethi Raghupatruni, Lurie, Chicago, IL, United States; Caren Mangarelli, Ann & Robert H. Lurie Children's Hospital of Chicago, Evanston, IL, United States; Brigid Gregg, C.S. Mott Children's Hospital, Ann Arbor, MI, United States; Daniel T. Robinson, Northwestern University The Feinberg School of Medicine, Chicago, IL, United States

Background: Diet quality in early childhood impacts growth and development. However, most data stem from children born at term, leaving sparse information on specific implications of dietary patterns for children born preterm (CBP).

Objective: To assess dietary patterns in CBP recorded at pediatric visits and evaluate associations between the patterns and growth and development.

Design/Methods: A retrospective cohort study included CBP ( < 37 weeks') with ³1 pediatric visit where dietary intake was recorded. Analysis included children ages ≥12 mo to 40 mo. Eligible visits occurred August 2023-August 2024, the first year our primary care visits used myplate.gov-based standardized dietary questions. For analysis, responses to 8 questions with ordinal or yes/no answers were converted to 13 dietary variables for principal component analysis and cluster analysis using imputation. Children were grouped as extreme (born 23-27 wks), moderate (28-33 wks) or late (34-36 wks) preterm. Growth and developmental milestones were recorded from any outpatient visit. Growth measures were corrected for prematurity through 24 mos. Linear and logistic regression models assessed association between dietary clusters (C, primary exposure) and growth and developmental milestones (outcomes), adjusting for preterm group using separate models for children with a single visit versus multiple visits.

Results: Among 148 CBP with 279 visits, two dietary C explained 47% of variance (Figure 1) with a small effect size of preterm group (Cramér's V=0.15). C1 included a higher proportion of children born moderate preterm with patterns of higher milk intake, less frequent dairy (cheese/yogurt) and more frequent "protein, not specified" responses. C2 included a higher rate of late preterm with lower milk, more dairy and more explicit protein choices (plant/fish/chicken/red meat). In single-visit analyses, C2 compared to C1 was significantly associated with higher weight Z score and height Z yet not with other single or multiple growth measures (Table 1). Children in dietary C2 had higher odds of meeting all developmental milestones (aOR=2.47, 95%CI:1.1,5.4). Among CBP not meeting all milestones, C2 was associated with lower odds of social/emotional developmental delay (aOR 0.24, 0.08,0.74) (Table 2).

Conclusion(s): Dietary patterns amongst CBP, as measured at primary care visits, suggest differential eating habits are not markedly different by degree of prematurity yet may be positively associated with growth and developmental milestones. This supports further investigation with more detailed diet assessment.

Figure 1.
Principal component analysis showing Clusters 1 and 2 and separation based on 13 factors which were derived from 8 individual questions with responses that were either ordinal (e.g., none, some, plenty) or yes/no as examples. 

Table 1.
 Linear regression results, separated based on whether children had one or multiple visits, estimating adjusted association between dietary clusters and growth outcomes. 

Table 2.
Cluster-robust logistic linear regression estimating association between dietary clusters and developmental assessments after adjusting for prematurity group.