The goal of this study, which describes a personal technique of continuous enteral nutrition (CEN) in hospitalized adults on an ambulatory basis, was: to prospectively evaluate, over a 2-year period, its efficacy and tolerance in 98 patients requiring CEN for at least 15 days; to compare its efficacy and tolerance with those of conventional non-ambulatory CEN on a prospectively randomized basis in 16 patients. Ambulatory CEN was given at the rate of 35-45 kcal/kg/d (lipids: 35 p. 100; carbohydrates: 45 p. 100); during day-time, a portable system, including pump, tubes and low-viscosity nutrient solutions, allowed ambulation. Ninety-eight consecutive patients with a minimal level of physical autonomy were treated for intestinal (n = 47), pancreatic (n = 20), esophagogastric (n = 17) diseases, or for malnutrition of other causes for an average of 38 days (15 to 141). The average weight gain (m +/- SD) was 1.2 +/- 5.5 p. 100 of ideal body weight (IBW) and the average nitrogen gain was 0.7 +/- 3.8 g/24 h; weight gain proved significantly lower in patients with inflammatory bowel disease receiving steroids. The clinical tolerance proved excellent, except for 5 cases of transient diarrhea and 9 cases of reposition of the nasogastric tube. A decrease in cholesterolemia below 3.9 mmol/l was noted in 25 p. 100 of patients during CEN. No significant difference between ambulatory and non-ambulatory CEN was observed in terms of evolution of body weight and other anthropometric variables, nitrogen balance, albuminemia, and oxygen consumption; conversely, the CEN experience, evaluated by patients on analogical visual scales, was significantly better endured in the ambulatory group.(ABSTRACT TRUNCATED AT 250 WORDS)