The incidence and causes of diarrhoea among critically ill patients receiving enteral tube feeding were investigated. Sixty acutely ill surgical or medical intensive care patients who had had a minimum of 48 hrs bowel rest were entered into the study. They were randomly assigned to receive one of two lactose free liquid formula diets - "Ensure", a commercially available feed containing 825 kCal/L and 34 g/L of protein with an osmolality of 441mOsm/1 or "Casilan Oil", a home-made feed containing 840 kCal /L and 45g/L of protein with an osmolality of 383 mOsm/1. The feeds were administered by constant nasogastric infusion. Patients received 1000ml at a rate of 40ml per hour for the first day and up to 2000ml at 80 ml per hour for the remainder of the study period. Investigations included documentation of medical history, medications administered and clinical details for each patient. Serum albumin was measured and the nutritional status of each patient was assessed using anthropometric measurements. Feeds were tested for bacterial contamination on the three days following the start of feeding and small intestinal bacterial overgrowth was assessed by the 1 g-¹⁴C Xylose breath test of Toskes and King. Twelve of the sixty patients had to be withdrawn from the trial within 24 hours of the start of enteral feeding for medical reasons. The remaining forty eight patients completed at least three days on enteral feeding and thereby became eligible for analysis. In 10/48 patients (21%) diarrhoea was present before enteral feeding began. Four of these 1 O patients continued to pass loose stools when enteral feeding was started while the remaining 6 settled. Diarrhoea developed in a further 10 patients (21%) after enteral feeding began. The overall incidence of diarrhoea in the group of critically ill patients studied was therefore 42% (20/48). However, of the fourteen patients who experienced diarrhoea during enteral feeding four had diarrhoea before feeding began. Therefore, the true incidence of diarrhoea related to enteral feeding was only 10/38 (26%). Furthermore, in 7 of these 10 patients, another possible cause of diarrhoea was present. There was no significant association between diarrhoea and nutritional status, hypoalbuminaemia, sepsis, length of bowel rest, sucralfate and antibiotic therapy other than amikacin. Twenty one patients received Ensure and 27 received Casilan Oil. Despite the differences in the composition of the feeds, the incidence of diarrhoea was similar on the Ensure and the Casilan Oil. No particular factor pertaining to the composition of the feeds was associated with diarrhoea. Significant contamination of feeds was universal but there was no constant relationship between bacterial counts, or types, and the occurrence of diarrhoea. Certain other factors were found to be significantly associated with diarrhoea. Abdominal injury was positively associated with the occurrence of diarrhoea (p<0.05). Diarrhoea could have been attributed to the underlying disease state in 7 of the patients. All three patients who were receiving lactulose as treatment for liver failure developed diarrhoea. While no association was noted between diarrhoea and antibiotic therapy in general, treatment with the antibiotic, amikacin, correlated significantly, albeit marginally, with the occurrence of diarrhoea (p<0.05). Twenty six patients were tested for small intestinal bacterial overgrowth. Only one patient, with an elevated excretion of ¹⁴CO₂, indicative of small intestinal bacterial overgrowth, developed diarrhoea. There was, however, a positive association between diarrhoea and decreased excretion of ¹⁴CO₂. It would appear that the bacterial flora was suppressed in patients with diarrhoea. Amikacin therapy was also associated with decreased excretion of ¹⁴CO₂. This may suggest that amikacin could have altered the bowel flora with resultant development of diarrhoea. While abdominal injury and disease were associated with the development of diarrhoea and amikacin was a possible factor associated with diarrhoea, the results of the present study indicate that enteral tube feeding with either the commercial feed, Ensure or the home-made feed, Casilan Oil was not a cause of diarrhoea in the majority of critically ill patients assessed. Furthermore, in most patients who commenced the trial with diarrhoea, improvement was noted on enteral feeding.