A prospective study was performed in a paediatric hospital to evaluate the incidence of bacterial contamination in enteral nutrition bags and to determine the critical points of process. During two separate one-month periods, all children receiving pump-assisted enteral nutrition were enrolled in the study. Samples for microbiological analysis were collected from enteral nutrition bags after administration in the first and second study period (sample T(2)). In the second study period, two additional samples were made at the end of the feed preparation process. One was refrigerated immediately (sample T(0)) and the other was sealed in a tube that followed the enteral nutrition solution until the end of its administration (sample T(1)). Bacterial contamination was detectable above 10(2)cfu/mL. Twenty-six out of 40 patients were included in the first study period and 14 out of 44 in the second study period. Contamination (>10(2)cfu/mL) occurred in nine of 26 samples (35%) and seven of 14 samples (50%) in the first and second study periods, respectively. Of these, five (20%) and three (21%) contained significant contamination (>/=10(4)cfu/mL). Bacteria of low pathogenicity were found in T(0) samples. Bacteria present in T(2) samples were pathogenic and multiple in 50% of cases. These results suggest that manipulation of the enteral nutrition bags at the bedside is critical for bacterial safety.