Developing dried milk spot based micro-sampling methods to assess changes to the fat composition of human milk during handling and processing

Abstract

Human milk (mother’s own milk and pasteurised donor human milk) is the food-of-choice of all infants, especially those born sick and/or prematurely. Premature infants might have to rely on nasogastric tube feeding of expressed and stored human milk rather than being fed directly on the breast before their suckling reflux matures. There has been increasing concern regarding whether the integrity of human milk is preserved after all the necessary storage, handling and processing steps involved in pasteurizing and delivering the milk to infants. Human milk fats are particularly of a concern as more than half of the energy of human milk is derived from milk fats. The endogenous lipases of human milk could breakdown the main fat type triglycerides (TG) and release free fatty acids (FFA), under various conditions. The main goal of the thesis was to develop micro-sampling based dried milk spot (DMS) methods to measure the fat composition of human milk. Compared to the conventional liquid milk analysis, collecting human milk as DMS reduces the volume required for analysis, requires only ambient temperature storage and transportation, and simplifies the analytical procedures. I first adapted the established dried blood technique for profiling the total fatty acid composition of human milk, using 200 milk samples of mothers from three different countries. The strong correlation and tight variation between human milk samples analysed using the conventional method and the DMS method gave me the confidence to move forward with DMS technique. The major challenges involved in the process was to develop a DMS method for measuring FFA concentration of human milk, due to the instability of milk fats and the difficulty in separating FFA from TG. I tested several strategies and was able to reduce the contamination from TG to a very low level of 2%, which however was unacceptable as this could result in a falsely inflate the reading of FFA. I then discovered that milk fats collected as DMS are susceptible to lipolysis due to breast milk lipases. Various attempts were made to inactivate the lipases. The final working system involved collecting human milk on silica gel impregnated paper, followed by microwaving to denature lipases. Milk fats can then be eluted and analysed by gas chromatography using an acid modified column that specifically detects FFA even in the presence of other fats (e.g. TG). The level of FFA measured by the DMS and conventional thin layer chromatography method were highly correlated (r=0.983, P<0.0001). To test the applicability and sensitivity of my DMS method, I then analysed 256 human milk samples collected in neonatal nursery at Women’s and Children’s Hospital, from a cohort of 32 mothers who delivered mostly preterm infants. In conclusion, this thesis is the first report of a DMS technique for reliably measuring the FFA concentration of human milk. This DMS technique can be adapted by human milk banks for monitoring the quality of milk fats throughout processing procedures, it also has the potential to be adapted for measurement of FFA in other biological fluids.