383 - The Association between Maternal Dietary Patterns and Early Preterm Milk Macronutrient Composition: A Secondary Analysis of the POGO Project Cohort.

Publication Number: 3370.383

Ruthfirst E. A.. Ayande, Yale School of Medicine, New Haven, CT, United States; Kristin Santoro, Harvard Medical School, Boston, MA, United States; Sarah Newell. Taylor, Yale School of Medicine, New Haven, CT, United States; Camilia R. Martin, Weill Cornell Medicine, New York, NY, United States

Background: Evidence linking maternal diet to preterm human milk (PTHM) composition is growing but remains mixed. Most studies focus on individual nutrients; however, examining overall diet quality may offer valuable insights into interventions to optimize enteral nutrition for preterm neonates.

Objective: To assess the association between maternal dietary patterns (DP) and the macronutrient composition of PTHM.

Design/Methods: Data were obtained from a multicenter cohort of very preterm infants (VPI) born < 32 weeks' gestational age (GA). Maternal milk samples, regularly collected and analyzed for macronutrients using mid-infrared spectroscopy, were included if expressed within the first 2 weeks post-birth (early PTHM). Using the Rapid Eating Assessment in Participants-Shortened version questionnaire administered in the first week, mothers provided information about their eating habits. Three latent class analysis (LCA) models with 2, 3, and 4 classes were built to identify DP, and model fit statistics were examined to determine the most suitable model. Generalized linear models were used to assess differences in milk composition between the identified DP from the LCA.

Results: Data from 267 participants were available for the LCA models. Two DP were identified: 1) a non-dietary guidance-driven pattern, NDGP (49.1%), and 2) a dietary guidance-driven pattern, DGDP (50.9%). The NDGP was driven by a high likelihood of sometimes skipping breakfast, consuming fewer than 2 servings each of whole grains, fiber, fruits, and vegetables; sometimes consuming processed foods, and sweets more than twice a day (Table 1). The DGDP was driven by a high likelihood of rarely skipping breakfast, rarely consuming fewer than two servings of whole grains, fruits, and vegetables, and rarely consuming processed food and soda (Table 1). Higher education (p <.0001) and low food insecurity (p=0.0005) were observed for the DGDP (Table 2). When adjusted for infant's birth GA, maternal education and food insecurity, the DGDP milk had higher true protein [Mean difference (SE) = 0.12 (0.07) g/dl, p = 0.06], and carbohydrate [Mean difference (SE) = 0.13 (0.14) g/dl, p = 0.36] concentrations, but these did not meet the conventional threshold for statistical significance (Fig. 1).

Conclusion(s): This is the first description of two DP in mothers of VPI, with higher food insecurity and lower maternal education relating to a NDGP. Differences in milk true protein and carbohydrate composition warrant further investigation, not only for nutritional impact but also for the potential immune impact of oligosaccharides and bioactive proteins.

Table 1: Conditional Probabilities (Rho Estimates) for Responses to REAPS Questions according to 2-Class LCA Model. Class 1: Non-dietary guidance-driven pattern and Class 2: Dietary guidance-driven Pattern. N=267
REAPS: Rapid Eating Assessment in Participants-Shortened version (REAPS) questionnaire;
LCA: Latent Class Analysis;
Class 1: Non-dietary guidance-driven pattern,
Class 2: Dietary guidance-driven pattern.
Item Response boldened when the probability of response is ~50% or more.


Table 2: Distribution of Infant and Maternal Characteristics by Dietary Patterns


Fig 1: Milk Macronutrient Content by Dietary Patterns.
Figure 1 shows the differences in mean macronutrient composition between the two identified dietary patterns. A: The dietary guidance-driven pattern had lower non-statistically significant Calories [Mean difference=-0.80, 95% -5.67, 3.96, p=0.74] compared to the non-dietary-guidance-driven pattern. B. Carbohydrates were higher in the dietary guidance-driven pattern compared to the non-dietary guidance-driven pattern [Mean difference=0.13, 95% CI:-0.15, 0.41, p=0.36], and this was not statistically significant. C: The dietary guidance-driven pattern had lower fat concentrations compared to the non-dietary guidance-driven pattern [Mean difference = -0.20, 95% CI: -0.66, 0.26, p = 0.39], and this was not statistically significant. D: Protein concentrations showed a statistical trend towards being higher in the dietary-guidance-driven pattern compared to the non-dietary guidance-driven pattern [Mean difference=0.12, 95% CI:-0.00, 0.25, p=0.06].