Concentrating Human Milk: An innovative point-of-care device designed to increase human milk feeding options for preterm infants

Abstract

Objective

The purpose of this study was to determine whether a point-of-care osmotic device concentrates important human milk (HM) nutrients to support feeding neonates requiring high-nutrient, low-volume feedings.

Study Design

Raw and pasteurized HM samples were concentrated to determine the effects of time and temperature on concentration. Concentrated samples were compared with matched baseline samples to measure changes in selected nutrient concentrations. Furthermore, changes in concentration of certain bioactive components of raw milk samples were measured.

Result

The device significantly increased the concentrations of the majority of the measured nutrient and bioactive levels (P<0.05). Increasing temperature of HM from 4°C to 37°C increased the concentration rate >30%. In all cases, the concentration rate of pasteurized HM was greater than that of raw HM.

Conclusion

The osmotic concentration of HM is a promising option for neonatal nutrition. Further studies are needed to establish an evidence base for the practical applications of this point-of-care device.

Introduction

Mother’s own milk provides the immune factors, stem cells, enzymes, and macronutrients essential for the optimal gastrointestinal development, immune function and growth of preterm infants.^1–4 However, lower birth weight preterm infants often cannot tolerate the large volume of human milk that would be required to meet their nutrient needs for adequate growth. In such cases, many neonatal intensive care units (NICUs) add bovine milk-sourced or human milk (HM) derived fortifiers to mother’s own milk or donor human milk to maximize the nutrient concentration of the milk. This supplementation is performed even when mothers have surplus milk supply because the nutrient density of HM varies and preterm infants usually require a higher concentration of nutrient intake by volume than term HM typically provides.^5–10

Feeding intolerance in preterm infants is associated with gut inflammation, sepsis, and necrotizing enterocolitis. Several studies have demonstrated an increased risk of necrotizing enterocolitis in infants receiving a bovine-based diet including bovine-derived HM fortifier.^11–15 In addition, milk feedings containing fortifiers derived from cow’s milk may change the microbiomes of preterm infants.¹⁶ The overall impact of such shifts in the microbiome is not entirely known, but changes in microbial diversity and abundance affect long-term health.^17–20 The protection of the gut microbiomes of preterm infants may be beneficial because of the medical risk associated with immature organs, environmental exposure, and stress. HM-derived fortifier use is limited because of its high cost and “limited efficacy data.”²¹

HM-derived fortifiers also displace intake of mother’s own milk. This displacement may impact the long-term health benefits of breastfeeding for both the mother and child because of the individualized benefits of mother’s own milk.

In many NICUs, donor human milk is available to supplement the mother’s own milk supply on an as-needed basis. Donor HM retains a significant amount of immunoprotective factors after the Holder pasteurization process and is associated with a decreased risk of necrotizing enterocolitis in comparison with bovine-derived formula feedings.^22–25 One potential drawback of donor HM is variability in nutrient density, which may be greater than that observed in mother’s own milk due to variation in timing of postpartum milk donation and loss of nutrients in milk processing. ^26–27 However, for the feeding of preterm infants, this variability affects the ability to achieve adequate nutrient intake by volume to meet the estimated nutrient needs of preterm infants.²⁸

A point-of-care process for concentrating HM provides a new option for achieving adequate nutrient density by volume for the feeding of preterm infants. This process also preserves nutrients usually diminished when using other fortification methods. The passive osmotic process that does not displace mother’s own milk nor does it require heat or external pressure to concentrate mother’s own milk, donor HM, or a combination of the two. The Human Milk Concentration device is comprised of a filtration membrane packet that uses osmotic draw to remove small amounts of water from human milk. The single use, sterilized device is made of food safe materials and fits into a standard sized milk storage container for use at point of care.

The objective of this study was to passively concentrate nutrients and bioactives in raw and pasteurized HM using an osmotic process to increase the nutrient content of a single human milk feeding. The study was designed to assess the Human Milk Concentration device’s concentration of previously frozen raw HM or pasteurized donor HM under temperature conditions common in NICU settings for HM storage and preparation of infant feeding.

Methods

Design

This study was considered exempt from institutional board review by the Human Milk Bank Association of North America’s research committee due to the use of rejected human milk donations that did not meet criteria for infant feeding. Permission to use this milk for research was granted by donors who signed the standard human milk donation agreement release form used by the Human Milk Bank Association of North America’s milk banks.

Water was removed from HM to increase the concentration of nutrients in both raw and pasteurized HM using the Human Milk Concentration device under room temperature (20– 23°C), refrigerated (4°C), and warmed (37°C) conditions. The temperature settings were similar to conditions mother’s own milk and donor human milk are stored in NICUs: fresh mother’s own milk kept at bedside/ room temperature (20–23°C) for feeding within hours of expression, standard fresh/ thawed mother’s own milk, donor human milk storage conditions are refrigerated (4°C), and mother’s own milk/ donor human milk are warmed prior to feeding (37°C). The water removal efficiency was defined as the rate at which water was removed from both raw and pasteurized HM samples over time. The fat, total protein, true protein, lactose, and non-protein nitrogen concentrations were measured before and after water removal to quantify differences in nutrient content associated with milk type (raw vs. pasteurized) and/or temperature.

The effectiveness of the device for concentrating bioactives was measured using raw milk concentrated for two hours at 4°C. This timeframe was selected because standard NICU feeding schedules are three to four hours apart so this would fit within NICU workflow. Nutrient concentration efficiency was compared between pre- and post- concentration contents of the milk, which included monitoring the following nutrients and bioactive components: fat; total protein; non-protein nitrogen; total protein; and amino acids; insulin; leptin; Immunoglobulins A (IgA), G (IgG), and M (IgM); epidermal growth factor.

Setting

HM concentration studies were performed at Northeast BioMedical in Tyngsboro, Massachusetts, between November 2017 and April 2018. As described in the protocol, pre- and post-concentrated HM samples were then shipped to the Proteomics and Metabolomics Facility at Colorado State University in Fort Collins, Colorado, and laboratories at the University of Colorado, Boulder, and Anschutz Medical Campus in Aurora, Colorado, for nutrient analysis.

Samples

The HM samples were obtained from Mothers’ Milk Bank Northeast (Newton, MA, USA) and provided because they did not meet the criteria for infant feeding set by the Human Milk Banking Association of North America. No information regarding the donor mothers or their infants was obtained. After thawing the HM samples, 60 mL aliquots were placed in 2.7-oz food contact grade polypropylene breast milk storage bottles. The Human Milk Concentration device was placed in the bottle with HM, and the bottle was capped and allowed to concentrate HM for 2 hours under the aforementioned temperature conditions.

Skim fractions of each HM sample were generated by centrifugation at 10,000×g for 10 minutes at 4°C and frozen at −20°C until nutrient and hormone concentrations were measured.

Measurement and data collection and analysis

The initial milk volume (HM) at t=0 (HM₀) and membrane weight (MEM_t) at t=0 (MEM₀) were measured separately. Changes in the membrane weight and milk volume were calculated for every 10-minute period. The difference between the initial membrane weight (MEM₀) and the membrane weight at time t (MEM_t) was representative of the amount of water absorbed by the membrane. The difference in the initial milk volume (HM₀) and the milk volume at time t (HM_t) was representative of the concentration process (water removal) of the HM. The experiment was terminated when the milk had been concentrated by at least 20%, i.e., [(HM₀-HM_t)/HM_{0] *}100%≥20%.

The initial milk volume at t=0 (HM₀) and membrane weight at t=0 (MEM₀) were measured separately. After a two-hour incubation (t₁₂₀), changes in the membrane weight and milk volume were calculated using the reported weights of the membrane and milk prior to and after milk concentration, MEM₁₂₀-MEM₀ and HM₁₂₀-HM₀, respectively. Subsequently, numbered samples were submitted for analysis (n=30 for analysis of macronutrients, insulin and leptin; n=20 for analysis of immunoglobulins at both t₀ and t₁₂₀). Once the nutrient concentration results were calculated, these values were compared to the concentration of milk, [(HM₀-HM₁₂₀)/HM_{0] *}100%.

Fat, protein, true protein, lactose, and non-protein nitrogen concentrations

To analyze the macronutrient content of HM, 2 mL aliquots of each sample from pre- and post-concentrated whole milk were diluted 10-fold with double-distilled H₂O. The diluted samples were warmed for 15 minutes to 38.5°C in a water bath prior to analysis. Triplicate measurements of milk macronutrients (fat, protein, true protein, lactose, and non-protein nitrogen) were obtained using a LactoScope Fourier transform mid-infrared spectrophotometer (Delta Instruments B.V., Drachten, the Netherlands).

Insulin, leptin, and immunoglobulin concentrations

Commercially available immunoassays were used to determine the pre- and post- nutrient concentrations of the following nutrients: insulin (chemiluminescent immunoassay; Beckman Coulter Access 2, Billerica, MA, USA); leptin (ELISA; ALPCO, Salem, NH, USA); IgA, IgG, and IgM (Bethyl Laboratories, Montgomery, TX); and epidermal growth factor, ThermoFisher Scientific, Waltham, MA).

Amino acid concentrations

High-performance liquid chromatography was used to measure the amino acid concentrations of the 30 skimmed samples, prior to and after HM concentration (pre- and post-free amino acid concentrations, respectively).

Statistical analysis

Data were compared using a two-sample t-test using Minitab 19. Statistical significance was defined as p<0.05. A two-sample t-test was conducted to quantify the water removed from the raw and pasteurized HM samples at each temperature to determine if the average water removal from the raw and pasteurized milk was significantly different. The time needed to obtain a 20% decrease in concentration was calculated by fitting each data set using a quadratic regression model in which y=0.2 for each condition (temperature and milk type). The time to 20% for each condition was then averaged to determine the time to 20% concentration. A one-way analysis of variance (ANOVA) test (α=0.05) was used to determine if there were statistically different average concentration times between conditions.

Results

The Human Milk Concentration device concentrated raw and pasteurized milk for all three test conditions, which included refrigerated, room temperature and warmed HM. Figure 1 shows a representative graph of the milk volume decrease over time for each experimental condition. Similarly, all samples showed an increase in membrane weight over time as the water passed through the membrane (Figure 2).

Figure 1.

Representative graph of pasteurized and raw milk volumes of refrigerated, room temperature, and warmed milk over time

Figure 2.

Representative graph of pasteurized and raw milk membrane weight for refrigerated, room temperature, and warmed milk over time

An average temperature-dependent volume reduction of 20% occurred between ~40 and 70 minutes for all experimental conditions (Table 1). There was no significant difference (p<0.05) between the time it took to achieve 20% concentration for warmed pasteurized and raw human milk. There was a significant difference (p<0.05) between raw and pasteurized milk concentration time at room temperature and refrigerated temperature. In each case, the pasteurized milk concentrated faster than that of raw milk.

Table 1.

Average number of minutes ±SD required to obtain a 20% increase in concentration for raw and pasteurized milk at different temperature conditions (n=sample size)

	Raw (n)	Pasteurized (n)
Refrigerated (4°C)	69.2±14.5 min (28)	55.6±14.8 min (30)
Room Temperature (20–23°C)	67.9±14.2 min (31)	54.5±9 min (30)
Warmed (37°C)	43.4±7.5 min (27)	40.3±6.3 min (30)

Variations were observed in each of the individual nutrient concentrations of 30 samples that had been concentrated for two hours using the Human Milk Concentration device (Figure 3). The average concentration of raw HM volume was 30±4%. The percent concentration of multiple milk components (fat, insulin, IgA, IgG, EGF, and IgM) did not differ significantly from the concentrated percent concentrations (Figure 3; Table 2). However, the average percent concentrations of several other milk components (protein, lactose, total protein, non-protein nitrogen, and leptin) varied significantly (p<0.05) when compared with that of HM.

Figure 3.

Boxplots of percent (%) change in milk concentration compared with nutrient concentration

Table 2.

Sample size (n) and average percent ±SD (%) concentration change of raw milk concentrated by the HMC device under refrigerated conditions. Asterisks denote nutrients that did not significantly differ from changes in milk volume (α=0.05).

Nutrient	n	% Change
Milk	30	30±4%
EGF*	20	34±30%
IgA*	20	31±26%
IgG*	20	41±21%
IgM*	20	77±71%
Insulin*	30	29±16%
Fat*	30	24±13%
Leptin	30	−3±27%
Protein	30	22±6%
Lactose	30	62±9%
True Protein	30	13±8%
Non-protein nitrogen	30	45±17%

The amino acid concentrations increased in all amino acids presented (Table 3). For all amino acid concentrations that increased after the HM concentration process, the average concentration ranged from 27% to 44%. As shown by the high standard deviations, there is a lot of variation observed.

Table 3.

Average percent ±SD (%) concentration change for raw milk and amino acids concentrated by the HMC device under refrigerated conditions

Nutrient	Percent Concentration
Milk	30±4%
Lysine	32±37%
Taurine	27±21%
Methionine	34±28%
Proline	45±58%
Phenylalanine	34±25%
Cysteine	35±16%
Aspartate	39±40%
Tyrosine	44±35%
Leucine	43±31%
Alanine	37±13%
Glycine	37±14%
Valine	37±15%
Threonine	39±13%
Histidine	40±17%
Serine	39±13%
a-Aminobutyric acid	41±18%
Isoleucine	43±31%
Citrulline	38±27%

Discussion

HM nutrients were significantly concentrated in and water was passively removed from HM samples via the osmotic Human Milk Concentration device, which resulted in increased macronutrient and bioactive levels overall. These results indicate this process may optimize HM feedings for preterm infants who require a higher amount of nutrients per feeding volume than term infants require.^29–33

The difference in time required to concentrate raw HM versus pasteurized HM observed in this study may have been because of changes in nutrient structure following pasteurization. Changes in the concentration of raw milk compared with that of pasteurized milk could have occurred because of differences in the osmotic pressure, solubility, and/or viscosity of the milk. However, this result should be further investigated as no difference was notable at a warming temperature of 37°C. For future use, the timeline of the change in volume using the Human Milk Concentration device may be decreased or adapted upon prototype refinement to best integrate the Human Milk Concentration device into NICU workflows.

HM macronutrients and several hormones remain stable during HM storage though enzymatic activity may be affected by length of storage time.³⁴ Given the inherent heterogeneity of HM, a larger number of controlled experiments, preferably performed with raw milk that did not undergo multiple freeze/thaw cycles, will be needed to determine if these findings are consistent over a larger sample size. The Human Milk Concentration device was designed for point of care use for HM that is either freshly pumped, stored in a refrigerator, or potentially while being warmed. HM concentrated by the Human Milk Concentration device would be anticipated to have an ‘expiration’ time similar to HM that has had a fortifier added to it (use within 24 hours).

Recent research comparing the growth and outcomes of feeding mother’s own milk compared with donor human milk supports increasing preterm infants’ intake of mother’s own milk to reduce risk of infection and feeding intolerance and increase growth rate.^35–38 Therefore, at the point of care, the option of concentrating either mother’s own milk only or a combination of mother’s own milk and donor human milk may be more advantageous than adding bovine-sourced fortifiers to mother’s own milk or donor human milk.

Feeding preterm infants concentrated mother’s own milk may be more beneficial than current supplementation methods because mother’s own milk contains higher levels of active enzymes, hormones, and other bioactives, including living cells such as stem cells. Because of variations in HM macronutrient compositions, further testing is needed to determine if the removal of a prescribed volume of water in earlier stages of lactation (such as preterm and early postpartum lactation) provides a similar result to fortifying mother’s own milk and donor human milk of varying nutrient density, which is the current standard practice in NICUs.

Future studies should address several limitations and results of this study. The lactation stage of the milk donors was not identified in this study. Since arginase activity is dependent on stage of lactation, with its activity declining as lactation progresses, the potential breakdown of arginine from varied storage conditions cannot be assessed by the present studies. ³⁹ Amino acid concentration changes noted in arginine and glutamine had high negative changes as well as high positive changes and the high-performance liquid chromatography peaks were not distinct. Due to this large range of results, the data for these two amino acids lacked statistical significance. That result, in combination with the assay selected for the detection of arginine and glutamine in human milk, demonstrates the quantification method may not have had sufficient sensitivity for determining concentration changes. Thus, in future studies the arginase activity of fresh or raw HM (samples from earlier lactation stages) should be examined prior to and following use of the Human Milk Concentration device.^40,41 The concurrent study of arginine levels in both fresh/raw and pasteurized HM samples at baseline (before concentration) and after concentration with the HMC device should be compared with baseline HM control samples to monitor the potential degradation of arginine. Glutamine is considered a sensitive amino acid that, depending on time and temperature handling, can partially degrade to glutamate over time. This phenomenon may have contributed to the large range in pre concentration levels and insignificant overall change observed in glutamine levels in the concentrated milk.^42–43 No nutritional inference can be made from the arginine and glutamine concentrations in these studies.

Lactose levels in the HM samples at baseline compared with lactose levels after concentration of the HM should be studied using a more sensitive method of analysis. Validated high performance anion exchange chromatography combined with pulsed amperometric detection method, may be useful in future studies.⁴⁴ Use of donor milk samples that were rejected by milk bank standards in this study meant that an unknown number of HM samples may have had prior contamination, improper storage, or high presence of bacteria so study of bacterial growth or contamination was not practical though this is recognized as important for future studies.

These proof-of-concept studies should be followed by studies with larger sample sizes to further investigate this novel point-of-care HM concentration process for feeding preterm infants. Osmolality is an important factor in assessing the process of HM concentration in comparison with other feeding options, and it is estimated that HM (approx. 300mOsm/kg at baseline) when HM is concentrated 20–30% would still be within recommended feeding ranges.⁴⁵ The Human Milk Concentration device process may be optimized for HM at lower temperatures, including use of the Human Milk Concentration device while fresh HM, or thawed mother’s own milk/ donor human milk is being stored while refrigerated, for later feeding. The process may also be optimized for use in fresh mother’s own milk stored at room temperature shortly after being pumped, or additionally be placed in mother’s own milk/ donor human milk while it is being warmed for feeding.

Future studies may elucidate additional benefits of point-of-care concentration of HM by this novel Human Milk Concentration device. Planned studies on fresh HM include: 1) measurement of enzymes, osmolality and stem cell levels in concentrated milk, 2) further micronutrient, lactose, bioactives and microbial growth testing of concentrated HM, and 3) optimal user protocol refinement needed for the FDA evaluation of the Human Milk Concentration device for use in feeding human infants. The purpose of these future studies is to demonstrate that this process preserves vital HM components that other fortification processes do not. Due to lack of precedent for this point-of-care method for HM concentration, future studies will also include FDA testing protocols as well as HM feeding preparation protocols in NICUs to assess how this method may be best incorporated into NICU staff workflow.

Conclusion

The results of this study show a point-of-care osmotic Human Milk Concentration device effectively concentrated HM, demonstrating a unique ability to optimize the use of mother’s own milk and donor human milk for preterm infant nutrition.