Osmolality of fortified human milk with contemporary human milk fortifiers

Abstract

Background

Human milk fortification is critical to support preterm infant growth. Human milk fortifed to caloric densities above manufacturers’ recommendations are often necessary to achieve adequate growth. Fortified milk may be prepared 12–24 h before feeding, although practice variations exist. We investigated how human milk osmolality is affected by contemporary bovine and human milk–derived human milk fortifiers (HMFs) immediately and temporally for standard and higher‐calorie recipes.

Methods

This study measured the osmolality of human milk fortified with four current HMFs across a range of caloric concentrations (22–30 kcal/oz) in fresh, previously frozen, and pasteurized human milk. Osmolality was tested at 0, 12, and 24 h after preparation, conducted in triplicate. Differences in initial osmolality were compared between fortifiers. Percentage change in osmolality at 12 vs 24 h was tested with mixed‐effects linear regression to study the effect/interaction of time, fortifier, and caloric density.

Results

Initial osmolality was significantly different between manufacturers (range, 315–620 mOsm/kg across caloric densities). There was a median 0.3% (IQR, −0.8 to 1.5) difference in change in osmolality at 24 vs 12 h (P = 0.02). In mixed‐effects modeling, there appeared to be an interaction between time and fortifier, but after post hoc adjustments, no pairwise comparisons remained significant. At 24 h, there was an interaction between fortifier and milk type (P = 0.003), with up to 4% greater change in osmolality with fresh vs frozen milk.

Conclusion

Contemporary HMFs increase the osmolality of human milk to variable degrees, with some temporal effects observed, although clinical significance of these findings is unclear.

BACKGROUND

Fortification of human milk is considered standard of care to optimize nutrition intake in preterm infants and promote growth in the neonatal intensive care unit. ¹ , ² , ³ Human milk fortifiers (HMFs) are commercially available products, derived from either human or bovine milk, designed to supplement human milk to achieve higher caloric densities and provide recommended macronutrients and micronutrients. ⁴ Under the assumption that human milk is estimated to average approximately 20 kcal/oz (67 kcal/dL), HMFs are most commonly manufactured to provide fortification up to an assumed 24 kcal/oz (81 kcal/dL; “standard fortification”). ⁵ , ⁶ , ⁷ , ⁸ However, manufacturer‐recommended fortification to 24 kcal/oz may not be enough, as the nutrition content of human milk varies and depends on many factors, such as stage of lactation and whether it is the mother's own milk or donor breast milk, and preterm infants may also have additional comorbidities that contribute to suboptimal growth. ⁹ , ¹⁰ , ¹¹ Consequently, off‐label recipes to achieve higher caloric densities (26 kcal/oz [88 kcal/dL] or higher) are used in clinical practice to achieve adequate growth in cases of growth faltering or restricted enteral feeding volumes. ¹¹

Clinically, there can be resistance to using higher‐calorie feedings, as past studies indicate that fortification of human milk increases the baseline osmolality. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ The osmolality of infant feedings has been a historical concern due to the 1976 recommendation from the American Academy of Pediatrics (AAP) against formula feedings exceeding 400 mOsm/L (~450 mOsm/kg), owing to a case series of necrotizing enterocolitis with hyperosmolar formula. ¹⁹ , ²⁰ It is unknown whether this is applicable to hyperosmolar feedings containing human milk. However, a systematic review concluded that there exists a lack of adverse effects with an osmolality up to 500 mOsm/kg reported in the literature. ²¹ Of note, studies of human milk osmolality after fortification with HMFs specifically available in the US, at either standard or higher caloric densities, have reported on now‐outdated formulations of HMFs, did not include all products, or had limitations with the type of milk used for recipes. ¹³ , ²² , ²³

Furthermore, fortified human milk is considered safe for feeding up to 24 h after preparation. ²⁴ Despite this recommendation from the Academy of Nutrition and Dietetics, variation exists in practice, with some centers fortifying milk before feeding and others preparing batches up to 12 or 24 h in advance. There is conflicting evidence whether refrigeration storage of fortified human milk significantly increases osmolality by 24 h. ¹³ , ¹⁵ , ²² , ²³

Our study aims to address these gaps by comprehensively evaluating the osmolality of human milk fortified with contemporary HMFs (both bovine and human milk–derived) available in the US market across both standard and higher‐calorie recipes. We also investigate the change in osmolality of fortified fresh, frozen, and pasteurized human milk at 12 and 24 h after preparation, a factor with potential implications for preparation protocols.

METHODS

This was a translational study using research human milk from multiple individuals dispensed by our Human Milk Research Biorepository, as approved by our institutional review board.

Three base types of human milk were tested: fresh (never frozen) unpasteurized maternal milk, previously frozen unpasteurized maternal milk, and pasteurized donor milk. Fresh milk, stored refrigerated, was prepared within 24 h of expression. Previously frozen milk and pasteurized donor milk were thawed under warm running water. The pasteurized donor human milk, obtained from the OhioHealth Mothers’ Milk Bank (Columbus, Ohio), had previously undergone standard Holder pasteurization.

The following HMFs from the United States market were used: Enfamil Liquid Human Milk Fortifier Standard Protein (EHMF‐SP; Mead Johnson Nutrition); Enfamil Liquid Human Milk Fortifier High Protein (EHMF‐HP; Mead Johnson Nutrition); Similac Human Milk Fortifier Extensively Hydrolyzed Protein Concentrated Liquid (SHMF; Abbott Nutrition); and Prolact+ H²MF (PHMF; Prolacta Bioscience). ⁶ , ⁷ , ⁸ , ²⁵ This selection represents the most current products from the three major manufacturers of HMFs in the US. The PHMF was prepared at 22, 24, 26, 28, and 30 kcal/oz (74, 81, 88, 95, 101 kcal/dL, respectively) following manufacturer recipes; PHMF at 22 kcal/oz was prepared using an off‐label recipe. EHMF‐SP, EHMF‐HP, and SHMF were prepared at 22 and 24 kcal/oz following manufacturers’ recipes and at 26 kcal/oz preparation using an off‐label recipe. For EHMF‐SP and EHMF‐HP, because additional off‐label recipes for 28 and 30 kcal/oz are provided by the manufacturer, these densities were also tested. However, because micronutrients may exceed expert‐recommended levels for preterm infants ²⁶ when concentrated HMFs are used beyond 26 kcal/oz, these higher caloric densities were not tested for SHMF. Recipes are detailed in the supplementary materials (Table S1).

Osmolality of fortified human milk was tested in triplicate for each preparation using a freezing‐point depression osmometer (Osmo1, Advanced Instruments). Single‐batch testing was primarily used to limit waste with fresh human milk and PHMF products; in cases where repeat testing was needed because of measurement errors or insufficient measurements, an additional batch was prepared. Testing was conducted within an hour of preparation and then at 12 and 24 h after preparation. Milk samples were refrigerated at 4°C between time points.

Paired Wilcoxon signed rank tests compared the percentage change in osmolality at 12 and 24 h. Multiple linear regression models compared individual fortifiers. Mixed‐effects linear regression models examined the effects and interactions of time, fortifier, and caloric density. Tukey‐Kramer adjustment was used for post hoc multiple comparisons. Statistical analysis was performed using SAS version 9.4 (SAS Institute Inc). P values <0.05 were considered significant.

RESULTS

Table 1 shows the initial osmolality measurements immediately after recipe preparation for all fortification combinations, and Table 2 compares absolute differences in osmolality between individual fortifiers at each caloric density. Initial osmolality was greatest with SHMF, then PHMF, and lower with both EHMF‐SP and EHMF‐HP.

Table 1.

Initial osmolality of fortified human milk after preparation.

		Osmolality, mOsm/kg
	Caloric density, kcal/oz	Fresh maternal	Frozen maternal	Pasteurized donor
Unfortified		297 ± 3	300.7 ± 9.5	302 ± 18.2
Enfamil SP	22	315 ± 7.8	323.7 ± 4.7	325.3 ± 10.7
	24	339.7 ± 12.9	343.3 ± 9.3	362.7 ± 17.4
	26	352.7 ± 0.6	350 ± 14.8	376 ± 21.7
	28	380.7 ± 5.5	370.7 ± 14.2	383.7 ± 15.6
	30	405.7 ± 3.8	403.7 ± 3.5	401 ± 7.9
Enfamil HP	22	317 ± 5.3	320 ± 2.6	309.3 ± 7.1
	24	334.3 ± 7.5	333 ± 2.6	330.3 ± 1.5
	26	329 ± 6.1	334.7 ± 0.6	343 ± 5.3
	28	352 ± 6.1	360.7 ± 3.5	355.3 ± 15.3
	30	384.7 ± 7.5	376.3 ± 5.7	385.3 ± 11.9
Similac	22	418.3 ± 9.7	428 ± 21.8	428.7 ± 2.1
	24	516.3 ± 5.9	531.3 ± 6	528.7 ± 0.6
	26	606 ± 22.7	619.7 ± 7.6	619.3 ± 0.6
Prolacta	22	345 ± 17.5	335.3 ± 16.2	349.3 ± 2.5
	24	404 ± 42.5	375.3 ± 37.1	381.3 ± 4.9
	26	388.3 ± 9.8	380 ± 4	404.3 ± 11.2
	28	399.7 ±± 4.7	416 ± 15.6	415 ± 15.4
	30	426.3 ± 18.8	429.7 ± 12.1	436.3 ± 12.4

Note: Mean ± SD. n = 3 for all recipe combinations.

Abbreviations: HP, high protein; SP, standard protein.

Table 2.

Difference in initial osmolality by fortifier type and caloric densities.

Caloric density, kcal/oz	Reference fortifier	Comparison fortifier	Estimated difference, mOsm/kg	95% CI	Adjusted P value
22	Enfamil SP	Enfamil HP	−5.9	−18.3 to 6.5	1
	Prolacta	Enfamil HP	−27.8	−40.2 to −15.4	0.003
	Prolacta	Enfamil SP	−21.9	−34.3 to −9.5	0.07
	Similac	Enfamil HP	−109.6	−121.9 to −97.2	<0.0001
	Similac	Enfamil SP	−103.7	−116.1 to −91.3	<0.0001
	Similac	Prolacta	−81.8	−94.2 to −69.4	<0.0001
24	Enfamil SP	Enfamil HP	−16.0	−28.4 to −3.6	0.51
	Prolacta	Enfamil HP	−54.3	−66.7 to −41.9	<0.0001
	Prolacta	Enfamil SP	−38.3	−50.7 to −25.9	<0.0001
	Similac	Enfamil HP	−192.9	−205.3 to −180.5	<0.0001
	Similac	Enfamil SP	−176.9	−189.3 to −164.5	<0.0001
	Similac	Prolacta	−138.6	−150.9 to −126.2	<0.0001
26	Enfamil SP	Enfamil HP	−24.0	−36.4 to −11.6	0.02
	Prolacta	Enfamil HP	−55.3	−67.7 to −42.9	<0.0001
	Prolacta	Enfamil SP	−31.3	−43.7 to −18.9	0.0003
	Similac	Enfamil HP	−279.4	−291.8 to −267.1	<0.0001
	Similac	Enfamil SP	−255.4	−267.8 to −243.1	<0.0001
	Similac	Prolacta	−224.1	−236.5 to −211.7	<0.0001
28	Enfamil SP	Enfamil HP	−22.3	−34.7 to −9.9	0.05
	Prolacta	Enfamil HP	−54.2	−66.6 to −41.8	<0.0001
	Prolacta	Enfamil SP	−31.9	−44.3 to −19.5	0.0002
30	Enfamil SP	Enfamil HP	−21.3	−33.7 to −8.9	0.08
	Prolacta	Enfamil HP	−48.7	−61.1 to −36.3	<0.0001
	Prolacta	Enfamil SP	−27.3	−39.7 to −14.9	0.004

Note: n = 3 for all recipe combinations. Negative estimate differences indicate that osmolality with the comparison fortifier is lower than with the reference fortifier.

Abbreviations: HP, high protein; SP, standard protein.

Table 3 depicts the percentage change in osmolality at 12 and 24 h. When considering all samples together, there was a median difference of 0.3% (IQR, −0.8 to 1.5) between percentage change at 24 vs 12 h (P = 0.02). When considering milk type subgroups, the median difference was 0.3% for fresh (IQR, −1.0 to 1.5; P = 0.22), 0.6% for frozen (IQR, −0.6 to 2; P = 0.04), and 0.3% for donor (IQR, −0.8 to 1.2; P = 0.53).

Table 3.

Change in osmolality of fortified human milk at 12 and 24 h.

		Fresh maternal		Frozen maternal		Pasteurized donor
	Caloric density, kcal/oz	Change (12 h), %	Change (24 h), %	Change (12 h), %	Change (24 h), %	Change (12 h), %	Change (24 h), %
Unfortified		1 ± 1.2	0.9 ± 1.7	−1.2 ± 1.5	0.5 ± 1.2	0.4 ± 0.9	0 ± 2.7
Enfamil SP	22	0.5 ± 0.9	1.4 ± 0.7	1.5 ± 2.7	3.4 ± 4.7	1 ± 0.9	0.9 ± 1.3
	24	1.2 ± 0.6	1.2 ± 1.3	−0.1 ± 1.2	0 ± 2.3	−0.7 ± 1.3	0.7 ±± 2
	26	2.5 ± 0.4	3.4 ± 1.9	0 ± 1.1	0.9 ± 0.7	−0.1 ± 0.4	0.6 ± 1.3
	28	1.4 ± 0.8	3 ± 1	0 ± 5.8	3.6 ± 1.1	−0.5 ± 3.3	0.5 ± 4.7
	30	1.8 ± 0.3	1.9 ± 1.7	1.3 ± 3.1	1.6 ± 1.3	0.2 ± 1.9	0.4 ± 2.3
Enfamil HP	22	−0.6 ± 1.9	0 ± 1.4	3.9 ± 4.8	3.1 ± 2.7	−0.1 ± 1.6	0.6 ± 2.3
	24	−0.8 ± 2.6	−1 ± 1.9	0.4 ± 0.7	1.2 ± 1.6	0.7 ± 0.9	0.3 ± 1.6
	26	1.5 ± 0.9	0.9 ± 0.6	1.4 ± 1.1	2.3 ± 1.4	0.4 ± 0.9	1.3 ± 1.2
	28	1 ± 2.2	0.5 ± 1.7	0.4 ± 0.3	2.3 ± 3.9	0 ± 0.7	−0.1 ± 0.6
	30	0.5 ± 0.9	−0.2 ± 1.7	0.3 ± 1.8	1.3 ± 1.2	−0.2 ± 0.7	−1 ± 0.7
Similac	22	1.2 ± 2.9	0.9 ± 2.8	−3.6 ± 5	−4 ± 4.6	0 ± 0.6	1.6 ± 0.8
	24	3.9 ± 0.9	6.1 ± 1.3	1.8 ± 1.3	3.3 ± 2.1	0.7 ± 0.9	1.5 ± 0.5
	26	1.9 ± 3	4.1 ± 3.8	3.5 ± 3.2	−0.3 ± 1.5	0.7 ± 0.2	2.4 ± 0.7
Prolact+	22	0.1 ± 1.3	0 ± 2.6	0.4 ± 4.3	1.6 ± 3.9	0.8 ± 1.9	−0.7 ± 2.3
	24	4.6 ± 5.9	−2.2 ± 3.4	2 ± 0.9	0.7 ± 1.6	0.6 ± 3.1	−1.1 ± 1.2
	26	2 ± 2.2	0 ± 2.5	−0.1 ± 2	−1 ± 0.8	−0.5 ± 1.1	−2.2 ± 1.1
	28	−0.3 ± 2.4	2.5 ± 1.8	−0.4 ± 1.2	0.4 ± 2.3	−0.3 ± 1.8	−1.3 ± 1.8
	30	−0.7 ± 4.8	0.1 ± 5.6	−0.2 ± 3.5	2.4 ± 2.7	2.8 ± 1.9	1.5 ± 4.5

Note: Mean ± SD. n = 3 for all recipe combinations.

Abbreviations: HP, high protein; SP, standard protein.

In mixed‐effects modeling, there was an interaction observed between time and fortifier (P = 0.02) but not between time and milk type (P = 0.81), nor was there a statistically significant three‐way interaction (P = 0.08). However, in the final mixed‐effects model, adjusting for fortifier and milk type and including the interaction between time and fortifier, time was not significantly related to change in osmolality (beta estimate, 0.6% [95% CI, −0.3 to 1.5]; P = 0.20). Despite the statistically significant interaction between time and fortifier, after post hoc Tukey‐Kramer adjustments, no pairwise comparisons remained significant.

Examining each time point separately, the change in osmolality at 12 h was not different by fortifier or milk type. However, for change in osmolality at 24 h, there was an interaction between fortifier and milk type (P = 0.003). In post hoc comparison, change in osmolality was 4% greater with fresh milk than with frozen when using SHMF (95% CI, 1.8–6.3; adjusted P = 0.02). With fresh milk, the change was 3.6% greater with SHMF than with either EHMF‐HP (95% CI, 1.7–5.6; adjusted P = 0.02) or PHMF (95% CI, 1.6–5.6; adjusted P = 0.02).

DISCUSSION

In this translational study characterizing contemporary use of HMFs available in the US, we found that bovine milk–derived and human milk–derived HMFs increased the osmolality of all types of human milk and to varying degrees. To our knowledge, our work represents the most current selection of HMFs in the US: EHMF‐SP and EHMF‐HP were introduced as a nonacidified liquid HMF in 2020, and the version of SHMF used in this study was updated in 2024. EHMF‐SP, EHMF‐HP, and PHMF have been previously tested in only one other study, although Pineda et al used only donor milk, specifically milk that had been processed by retort sterilization and not the more commonly used Holder pasteurization method. ²³ , ²⁷ Two other groups have evaluated previous HMF iterations using either maternal milk alone ²² or pasteurized donor milk and frozen maternal milk. ¹³ By testing recipe combinations with fresh maternal, frozen maternal, and pasteurized donor milk, our study differed methodologically from these prior works.

Our finding that SHMF yielded the highest osmolality across all milk types, and frequently >450 mOsm/kg, is consistent with studies comparing previous formulations. ¹³ , ²² , ²³ This may be due to the differing nutrient composition of SHMF, which contains extensively hydrolyzed bovine milk casein and might confer a greater proportion to the initial osmolality shift. ⁶ Although the AAP threshold is often cited by clinicians who choose to refrain from using higher caloric densities, the theoretical risk for necrotizing enterocolitis with an osmolality above this cutoff has not been established with human milk or modern HMFs. Addition or administration of common enteral supplements, such as multivitamins or iron, alone can exceed this theoretical threshold. ²⁸ , ²⁹ Moreover, one potential trade‐off is that, unlike HMF with intact protein, HMF with extensively hydrolyzed protein does not increase fecal calprotectin levels, a marker of intestinal inflammation. ³⁰ , ³¹ Recipes using other products to achieve higher caloric densities can also exceed 450 mOsm/kg. ²² , ²³ Collectively with the growing literature that these recipes and products are clinically used and tolerated, especially with pasteurized donor milk, ³² , ³³ , ³⁴ our results add to the compelling argument that a thorough reevaluation and investigation of this benchmark in the context of contemporary fortifier compositions and modern neonatology is indicated.

We also reported minor temporal differences that were identified as statistically significant, including an interaction between fortifier and milk type at 24 h, although the clinical significance of these small changes is not clear. Fresh human milk is enzymatically rich, containing active components such as amylase, which likely contributes to observed changes via starch degradation, ¹² as SHMF contains a higher proportion of carbohydrate content compared with the other fortifiers. ⁶ , ⁷ , ⁸ , ²⁵ It is possible that our study was underpowered to appreciate the interaction between time and fortifier after multiple‐comparisons adjustment, but the slight difference in osmolality change at 12 vs 24 h may also not be clinically relevant. This discrepancy is underscored by the conflicting results from existing studies with previous iterations of US HMFs: Two reported no change over time, but Donovan et al observed a change of 11.6 mOsm/kg over 24 h. ¹³ , ²² , ²³ European studies with different HMF products have also reported a similar change in osmolality by 24 h, ¹⁴ , ¹⁵ , ¹⁶ with most of the increase occurring shortly after fortification. ¹⁷ , ¹⁸ Again, it is unknown whether these changes by approximately 3%–4% (also similar to our detected interactions between SHMF and milk type at 24 h) are clinically significant; however, many authors posit that it is likely to be negligible. ¹³ , ¹⁵ , ¹⁶ , ²² , ³⁵ In absence of established impact on clinical outcomes, advocating for 12‐h preparations over 24‐h preparations should not come at the expense of centralized milk preparation, which should still be prioritized. ³⁶

A major strength of our study is the broad range of milk bases and HMFs we measured, as well as the inclusion of an interim 12‐h measurement of osmolality to allow comparison of 12‐ and 24‐h preparations. Limitations of the study include potential delays in timing of osmolality measurements and the small number of replicates. Because the clinical laboratory operates the osmometer at our institution, recipes were prepared in our metabolic kitchen; then, samples were delivered to the laboratory technician. Owing to the number of recipe combinations and the logistical constraints related to the clinical laboratory, we were limited in the quantity of batches we could perform at one time. Thus, we opted for triplicate testing, and most of the testing was done on three samples from a single batch of prepared product. Additionally, research human milk was donated by multiple individuals with excess supply. Human milk properties can vary depending on the person as well as both postnatal and gestational age of the child, ¹⁰ and this may contribute to some variability between recipes tested. We also acknowledge that many of the higher‐calorie recipes are outside of manufacturers’ directions, and whereas our recipes achieved these densities by altering the ratio of liquid HMF to human milk, there are other ways to increase caloric density, such as adding formula powder, which we did not test.

In conclusion, our investigation into the osmolality profiles of fortified human milk provides insight on contemporary US fortifier products added to three common types of human milk, with both standard and increased fortification densities. Future research should focus on large‐scale, prospective studies to determine evidence‐based guidelines on optimal growth and nutrition with current human milk forms and fortifiers that include osmolality of infant feedings to fully inform the overall impact on the risk of necrotizing enterocolitis and other adverse and long‐term outcomes.

AUTHOR CONTRIBUTIONS

Abigail Gardiner contributed to data acquisition, data interpretation, and drafting of the manuscript. Irmalís Flores González and Carrie Smith contributed to conception and design of the research, data acquisition, data interpretation, and drafting of the manuscript. Jacqueline Wessel and Jae H. Kim contributed to the conception of the research. Ting Ting Fu contributed to conception and design of the research, data analysis, data interpretation, and drafting of the manuscript. All authors critically revised the manuscript, agree to be fully accountable for ensuring the integrity and accuracy of the work, and read and approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

Irmalís Flores González has received speaking compensation from Reckitt/Mead Johnson Nutrition. Jae H. Kim is a consultant for Medela, Infant Biotherapeutics, Mother's Milk Is Best, Biomilq; has a sponsored research grant with Fresenius Kabi; is a shareholder for Astarte Medical and Nicolette; and is an unpaid Scientific Advisor to the NEC Society. Ting Ting Fu has a speaking agreement with Medela. The research was conducted independent of these relationships. The remaining authors declare no conflicts of interest.

Supporting information

Table S1.