Calibration Adjustment of the Mid-infrared Analyzer for an Accurate Determination of the Macronutrient Composition of Human Milk

Background: Human milk composition analysis seems essential to adapt human milk fortification for preterm neonates. The Miris human milk analyzer (HMA), based on mid-infrared methodology, is convenient for a unique determination of macronutrients. However, HMA measurements are not totally comparable with reference methods (RMs). Objective: The primary aim of this study was to compare HMA results with results from biochemical RMs for a large range of protein, fat, and carbohydrate contents and to establish a calibration adjustment. Methods: Human milk was fractionated in protein, fat, and skim milk by covering large ranges of protein (0-3 g/100 mL), fat (0-8 g/100 mL), and carbohydrate (5-8 g/100 mL). For each macronutrient, a calibration curve was plotted by linear regression using measurements obtained using HMA and RMs. Results: For fat, 53 measurements were performed, and the linear regression equation was HMA = 0.79RM + 0.28 (R-2 = 0.92). For true protein (29 measurements), the linear regression equation was HMA = 0.9RM + 0.23 (R-2 = 0.98). For carbohydrate (15 measurements), the linear regression equation was HMA = 0.59RM + 1.86 (R-2 = 0.95). A homogenization step with a disruptor coupled to a sonication step was necessary to obtain better accuracy of the measurements. Good repeatability (coefficient of variation < 7%) and reproducibility (coefficient of variation < 17%) were obtained after calibration adjustment. Conclusion: New calibration curves were developed for the Miris HMA, allowing accurate measurements in large ranges of macronutrient content. This is necessary for reliable use of this device in individualizing nutrition for preterm newborns.