The use of sterile, prepackaged enteral feeds has helped to reduce the risk of exogenous microbial contamination of enteral feeds but there is increasing evidence that endogenous contamination with bacteria from the patient's own flora may occur. The purpose of this study was to compare the levels and types of micro‐organisms present in residual feed in nutrient containers and giving sets when either 500 or 1000 ml prefilled, ready‐to‐hang nutrient containers were used to administer 1000 and 2000 ml quantities of feed to patients on hospital wards over 24 h using a single giving set over this period. Forty‐one adult patients were randomly allocated to receive either 1000 or 2000 ml of sterile undiluted, whole protein feed over 24 h from 1000 ml prefilled containers (Nutrison Steriflo, Cow and Gate Nutricia Ltd) or from 500 ml prefilled bottles (Osmolite, Abbott Laboratories Ltd). One giving set was used over 24 h for each patient. Samples of feed from the nutrient containers and the distal end of each giving set were sent for microbiological analysis immediately after removal from the patient. Control experiments demonstrated that there were no micro‐organisms in the unopened feed containers and that none was introduced during the feed sampling procedure. The percentage of days on which feed samples from the nutrient containers were contaminated was $lt; 10% for feeds administered at a rate of 1000 ml over 24 h and $lt; 23% for those administered at a rate of 2000 ml over 24 h irrespective of system, which may be attributed to increased handling of systems and reuse of the giving set. However, the percentage of days on which samples from the giving sets were contaminated was 43% for all systems. In fact, on an average of 33% of days it was found that the only feed sample that was contaminated was that collected from the giving set. Counts for these samples were $lt;103 cfu/ml and the main bacteria isolated included Klebsiella spp., Enterobacter spp., Escherichia coli, Streptococcus faecalis and Pseudomonas aeruginosa. These results further implicate retrograde spread of the patient's own flora as a possible source of microbial contamination of feed in the giving sets of enteral feeding systems.