Abstract
Expressed breast milk used to feed preterm infants is precious and so, despite heterogeneity of composition, all available milk is used. A study of 274 samples of expressed breast milk supplied by preterm mothers and National Childbirth Trust donors showed pronounced variation in fat content as measured by the "creamatocrit" method. This was not due to differences between term and preterm mothers or between transitional and mature milk. The composition was affected by diurnal variation and method of collection. Substantial amounts of fat were also wasted as a result of continuous nasogastric feeding. Several milk samples did not contain enough fat to supply even a fraction of the recommended energy requirements of these infants. Some type of quality control over samples of expressed breast milk is clearly essential. The creamatocrit method is simple and feasible.