This randomised double-blind study examined the frequency of diarrhoea in intensive care unit patients given a fibre-containing feed, Enrich, and a fibre-free feed, Ensure. A daily 'diarrhoea score' was calculated from the frequency, volume and consistency of the stools. A score greater than 12 indicated diarrhoea, and greater than 50 severe diarrhoea. Ninety one patients met the criteria for participation: Enrich 45, Ensure 46. The groups were similar in age, sex ratio, feed volume, antibiotic usage, upper gastro-intestinal bleeding prophylaxis and plasma albumin concentrations. Diarrhoea occurred in 16 patients given Enrich and 13 given Ensure, severe diarrhoea affecting 5 and 9 respectively (NS). Forty seven of 343 (14%) Enrich feeding days and 51 of 342 (15%) Ensure feeding days were complicated by diarrhoea--severe diarrhoea: 8 and 12 feeding days (NS). We conclude soy polysaccharide (21 g/L) did not reduce diarrhoea in intensive care unit patients given enteral feeds.