Evaluating the Effect of Random Multi‐Donor Pooling on the Nutritional Variability in Donor Human Milk Using Computer Modeling

Protein and fat concentrations in donor human milk (DHM) can vary twofold to threefold and are influenced by the number of unique donors per pool. The aim of this study was to broadly characterize how the number of donors (2-10) randomly combined into a pool during milk bank processing influenced the variability of macronutrients, vitamins, minerals, and bioactive factors in DHM. The minimum number of donors required for 80% of the pools to meet pre-defined targets for true protein, fat, and disialyllacto-N-tetraose (DSLNT) was also evaluated. Monte Carlo simulation was used to create models that accounted for donor lifetime donation volume and milk bank production constraints. Variability in nutrients was quantified as a Nutrient Inequality Index (NII) which was computed as the ratio of the 90th percentile to the 10th percentile for each simulation. Random multi-donor pooling of 2-10 donors produced lower variability in DHM macronutrients than most vitamins and minerals. A priori targets of 0.9 g/dL of true protein, 3.5 g/dL of fat, and 210 µg/L of DSLNT could not be achieved with any random pooling scenario. The NII for lactose stabilized at less than 1.1 when there were 3+ donors per pool, while the NII for fat and true protein stabilized at less than 1.3 when there were 5+ donors per pool. The NII exceeded 1.5, even at 10 donors per pool, for several micronutrients including zinc, copper, sodium, iron, biotin, riboflavin, B6, B12, and pantothenic acid.