Infrared analyzers for breast milk analysis: fat levels can influence the accuracy of protein measurements

Background: Currently, there is a growing interest in lacto-engineering in the neonatal intensive care unit, using infrared milk analyzers to rapidly measure the macronutrient content in breast milk before processing and feeding it to preterm infants. However, there is an overlap in the spectral information of different macronutrients, so they can potentially impact the robustness of the measurement. In this study, we investigate whether the measurement of protein is dependent on the levels of fat present while using an infrared milk analyzer. Methods: Breast milk samples (n = 25) were measured for fat and protein content before and after being completely defatted by centrifugation, using chemical reference methods and near-infrared milk analyzer (Unity Spectra-Star) with two different calibration algorithms provided by the manufacturer (released 2009 and 2015). Results: While the protein content remained unchanged, as measured by elemental analysis, measurements by infrared milk analyzer show a difference in protein measurements dependent on fat content; high fat content can lead to falsely high protein content. This difference is less pronounced when measured using the more recent calibration algorithm. Conclusions: Milk analyzer users must be cautious of their devices' measurements, especially if they are changing the matrix of breast milk using more advanced lacto-engineering.