Karen Harris

Karen Harris is a professional writer and editor with more than two decades of experience in research-driven nonfiction writing. With a background in journalism and a Master's degree in English, she brings a meticulous approach to fact-finding, source evaluation, and clear communication of complex topics. Her work spans health, lifestyle, and consumer education, always with a focus on accuracy and readability.

Context

Preterm and low birthweight infants require energy-dense enteral nutrition, which means fortified human milk or preterm formula - both of which have higher osmolality than unfortified human milk. It has long been assumed that high feed osmolality increases the risk of adverse gastrointestinal outcomes, particularly feeding intolerance and necrotising enterocolitis (NEC). Despite this widespread clinical concern, the only existing recommendation - a 1976 American Academy of Pediatrics guidance capping formula osmolarity at 400 mOsm/L for normal infants - has never been substantiated by rigorous trial evidence. This is the first systematic review to examine the link between feed osmolality and adverse outcomes across human and animal studies.

Study Overview

Design: Systematic review (PRISMA); six databases searched from inception to May 2018; MEDLINE, Embase, CAB Abstracts, Current Contents, BIOSIS Previews, SciSearch


Included studies: 10 human studies (8 RCTs, 2 observational) + 6 animal RCTs


Population: Human: 618 newborn and low birthweight infants (predominantly preterm); Animal: neonatal dogs, rats, mice, piglets


Follow-up: Variable; most human studies assessed outcomes during acute feeding periods


Funding: No specific grant declared; three authors (Z-ME, HSGT, RMvE) are employees of Danone Nutricia Research


Evidence Certainty (GRADE): Not formally assessed; overall evidence quality low - small sample sizes, variable blinding, frequent differences in formula composition alongside osmolality, limiting causal interpretation


Comparison:

  • Feeds ranging from ~155 mOsm/kg (half-strength human milk) to 849 mOsm/kg (experimental hyperosmolar formulas) across human and animal studies
  • Key clinical range of interest: 300–500 mOsm/kg (covering fortified human milk and preterm formulas in routine use)

Key Findings

Gastric emptying (5 human studies):

  • No significant difference in gastric emptying across osmolalities up to 448 mOsm/kg in four studies
  • One study found delayed gastric emptying with a hydrolysed protein formula at 539 mOsm/L - but that formula also differed substantially in protein, fat, and carbohydrate composition, making osmolality the independent cause impossible to confirm
  • A follow-up study using matched compositions found no osmolality effect up to 448 mOsm/kg

Feeding intolerance (4 human studies):

  • No significant differences in feeding tolerance (abdominal distension, vomiting, delayed feeds) with osmolalities up to 451 mOsm/kg across all four studies

NEC incidence (5 human studies):

  • No significant difference in NEC with osmolalities up to 450 mOsm/kg in three studies
  • One small RCT (n=16) found higher NEC incidence with an elemental formula at 650 mOsm/L vs. cow milk formula at 359 mOsm/L (87.5% vs. 25%) - but the formulas differed in protein source, fat type, and carbohydrate composition; osmolality as the independent cause cannot be established
  • One retrospective study (Thoene et al.) found the highest NEC incidence with the lowest osmolality feed (326 mOsm/kg, acidified liquid HMF) - directly contradicting the assumed osmolality–NEC relationship

Animal studies:

  • Delayed gastric emptying and increased intestinal water found at osmolalities ≥624 mOsm/L in rats - well above the clinical range
  • Decreased survival due to dehydration in neonatal rats at ≥765 mOsm/L - not directly extrapolatable to humans, but raises concern for very high osmolality feeds
  • No significant differences in NEC incidence with varying osmolalities up to 849 mOsm/kg in animal NEC models
  • Multiple structural, functional, and immune intestinal differences observed in preterm piglets fed higher osmolality fortified feeds - though feed composition also differed

Overall pattern:

  • The only consistent signal of harm appears at osmolalities substantially above the clinical range (>539–624 mOsm/L)
  • In the routine clinical range of 300–500 mOsm/kg, no consistent evidence of harm was found across human studies
  • In most studies, formula composition changed alongside osmolality - making it impossible to isolate osmolality as the independent variable

Limitations

  • Three of five authors are employees of Danone Nutricia Research, a major formula manufacturer - a significant conflict of interest for a review of formula safety
  • Only one reviewer screened initial titles and abstracts, increasing risk of missed studies
  • Review restricted to full-text, English-language publications only
  • Feed osmolality was not measured in all studies - in some it was assumed, reducing reliability of cut-off conclusions
  • Osmolality was rarely the sole varying factor: most studies compared feeds that also differed in protein source, carbohydrate type, fat content, or fortifier type - making osmolality-specific conclusions impossible
  • Human studies were small (total n=618 across 10 studies); none was powered to detect differences in NEC incidence
  • Future RCTs would require >1,000 infants to be adequately powered for NEC endpoints
  • Animal-to-human extrapolation is uncertain; neonatal animal gut physiology differs from human preterm infants

Neutral Interpretation

This systematic review found no consistent evidence that milk feed osmolality in the range of 300–500 mOsm/kg - the range encountered in routine fortified feeding of preterm infants - is associated with adverse gastrointestinal outcomes including feeding intolerance or NEC. However, because nearly all studies varied formula composition alongside osmolality, the independent effect of osmolality cannot be established from existing data. The finding that the highest NEC incidence in one study occurred with the lowest osmolality feed further undermines a simple osmolality–harm relationship. Given that three of the five authors are affiliated with Danone Nutricia Research, independent replication of these conclusions is warranted before they are used to liberalize osmolality thresholds in vulnerable preterm populations.

Full Citation

Ellis ZM, Tan HSG, Embleton ND, Sangild PT, van Elburg RM. Milk feed osmolality and adverse events in newborn infants and animals: a systematic review. Archives of Disease in Childhood — Fetal and Neonatal Edition. 2019;104(5):F474–F481. doi: 10.1136/archdischild-2018-315946. 

PMID: 30523072. 

Available at: https://pubmed.ncbi.nlm.nih.gov/30523072/

Disclosure

This summary is based on the published systematic review. It is provided for informational purposes only and does not constitute medical advice. Funding sources and conflicts of interest are disclosed as reported by the authors.