Contact us / Members' area    

Home > Abstracts > Index > 2013 > Cooper A, et al


A Comparison between a Human Milk Analyser and Established Laboratory Methods for Measurement of Human Milk Macronutrient Content

Presented as a poster at the Neonatal Society 2013 Summer Meeting (programme).

Cooper A1, Dorrian C2, Barr M2

1 Neonatal Intensive Care Unit, Royal Hospital for Sick Children, Glasgow, UK
2 Department of Clinical Biochemistry, Royal Hospital for Sick Children, Glasgow, UK

Background: The nutrient content of mother’s own milk is known to be variable which poses difficulties in the context of feeding pre term and sick infants. Use of a human milk analyser (Miris AB, Sweden) as an aid to ‘tailored’ fortification of human milk has been proposed by several authors (1,2). The aim of this study was to compare the Miris Human Milk Analyser (HMA) with established laboratory methods used to measure macronutrient content.

Methods: 40 samples of donor expressed human milk were obtained within the regional donor milk bank. Each sample was analysed using the HMA to obtain macronutrient values for fat, protein, lactose and calculated energy content. In the laboratory, the Lowry method was used to measure milk protein. The Abbott Architect Analyser was used to measure milk fat. Lactose content was derived from a modification of the method used by Kuhn & Lowenstein combined with automated analysis. As with the HMA, energy values were calculated.

Results: Measurements for fat and lactose showed excellent correlation between the two methods with correlation coefficients of 0.97 and 0.96 respectively. Lactose measured by the HMA was slightly lower than by the laboratory method (mean difference = -0.3g/dL). One sample produced markedly discrepant results for lactose by the two methods and so was excluded from the correlation analysis. Fat measured by the HMA was slightly higher than the laboratory method (mean difference = +0.12g/dL). Protein measurements showed a slightly poorer correlation (r2 0.85), with the HMA producing slightly lower results than the laboratory method (mean difference -0.31g/dL). For calculated energy, correlation coefficient was 0.92. In this study interassay coefficient of variance (%CV) for the HMA was, 8.1% for protein, 2.0% for fat, 1.8% for lactose and 1.7% for calculated energy content (n=7).

Conclusion: The results show a good agreement between the Miris HMA and the standard laboratory methods for all macronutrient parameters. This makes the HMA a desirable option for rapid assessment of milk nutrient content in the clinical environment. The potential impact on neonatal nutrition must be further explored.

Corresponding author:

1. Casadio Y et al. Evaluation of a Mid-Infrared Analyser for the Determination of the Macronutrient Composition of Human Milk. J Hum Lact. (2010) 26(4) 376-83
2. Menjo A et al. Bedside Analysis of Human Milk for Adjustable Nutrition Strategy. Acta Paed (2009) 98, 380-84

Home / The Neonatal Society, founded 1959 / © 2001 - 2016