333 - Nutritional Gaps in Breastmilk Delivery and Growth Outcomes for Very Preterm Infants

Publication Number: 3321.333

Nicole C. Weitzel, Weill Cornell Medicine, New York, NY, United States; Liana Senaldi, Weill Cornell Medicine, New York, NY, United States; David Liu, Weill Cornell Medicine, Brooklyn, NY, United States; Divya Mandalaywala, Weill Cornell Medicine, New York, NY, United States; Camilia R. Martin, Weill Cornell Medicine, New York, NY, United States

Background: NICU breastmilk feeding protocols often assume standard macronutrient concentrations in human milk, though actual composition varies. The MIRIS Human Milk Analyzer measures exact nutrient content but is resource intensive. Understanding variation between assumed and actual nutrient values, and impact on growth, may support identifying when individualized breastmilk fortification is needed.

Objective: To determine how potential discrepancies between assumed and measured nutritional content of breastmilk impact growth to inform the role of personalized milk analysis for very preterm infants.

Design/Methods: Infants born at < 32 weeks gestation and receiving ≥ 20 mL/kg/day of breastmilk at the NYP-ACH NICU were included (n=35). Breastmilk aliquots were collected daily, pooled weekly, and analyzed via MIRIS for protein, fat, carbohydrate and caloric content. Donor milk was also analyzed. A total of 198 milk analyses were performed. Actual intake was calculated and compared to assumed intake. Linear regressions were conducted comparing discrepancies in actual and assumed nutrient intake against weekly growth velocity and z-scores, as well as growth metrics at day of life 30 and 36 weeks' postmenstrual age. Gestational age was controlled for in these analyses.

Results: Actual macronutrient and energy intake exceeded assumed intake in most weeks of breastmilk feeding (Figure 1). However, substantial variation existed within the cohort, especially for energy delivery. During the first four weeks of breastmilk feeds, between 19-39% of infants received less calories than assumed (Figure 2). Linear regression analyses revealed significant associations between weekly differences in actual and assumed nutrient delivery and key growth metrics throughout NICU stay. In particular, during the second month of breastmilk feedings, several regressions suggested that overestimation of intake, or a greater assumed intake than actual intake, was associated with poorer growth outcomes in the subsequent week (Figure 3).

Conclusion(s): The impact of the differences between actual and assumed nutritional delivery appeared greater during the later convalescent phases of feeding and nutrition. These findings suggest that individualized target fortification may be of greatest use later in the NICU course. However, given the variability in breastmilk composition and energy delivery throughout the entire NICU stay, earlier target fortification may benefit very preterm infants that experience postnatal growth failure.

Figure 1. Percent difference between average actual and assumed delivery of macronutrients and energy over time.
The percent difference was calculated as (actual value- assumed value)/assumed value X 100%). Positive values indicate greater actual delivery, and negative values indicate that the assumed content was greater than the actual content, resulting in nutrient under-delivery. *p < 0.05, **p < 0.01, *** p < 0.001

Figure 2. Individual variation over time between assumed and actual energy delivery.
Energy percent difference was calculated using the percent difference formulation described in Figure 1. An energy percent difference <0 indicates the infant received less calories from breastmilk than anticipated. Individual infants are represented with black dots, while the average energy percent difference for that week of breastmilk feeds is represented by the red diamond.

Table 1. Summary of significant associations between nutrition delivery differences and growth metrics.
Weekly nutrient differences were compared with weekly growth velocity (GV) and weight z-scores across weeks of breastmilk feeding using linear regression. Analyses that examined nutrient difference in the week prior compared weekly nutrient differences to growth metrics in the following week. Nutrient difference was defined as actual intake minus assumed intake, with positive values indicating greater nutrient delivery than assumed. Values represent β coefficients and 95% confidence intervals (CI). Only significant results are shown.