A controlled trial was performed to compare enteral feeding with either an amino acid based feed or a whole protein feed as sole treatment for active Crohn's disease. Twenty four patients were studied (nine with ileal, 11 with ileocolonic, and four with colonic disease). Both feeds proved effective; nine of 13 patients randomised to receive the amino acid based feed were in clinical remission within three weeks as defined by a simple activity index compared with eight of 11 treated with the whole protein feed. Patients in clinical remission were then crossed over onto the other feed. None of the six patients who were changed to the whole protein feed relapsed over the subsequent three week period compared with three of seven patients who were changed to the amino acid based feed. In responders the median serum C reactive protein concentration fell from 21 mg/l (range 9-82) on entry to 6 mg/l (range 3-19) at six weeks. Seven patients relapsed within eight months of starting solid food (mean 3.7 months), while nine were still in remission (follow up period 3-9 months, median six months). Detailed studies of staged reintroduction of food and permitted food additives were carried out over a four year period in a patient with extensive stricturing small bowel Crohn's disease who had been brought into remission by open treatment with enteral feeding. Carrageenan, other permitted emulsifiers, bread, meat, potatoes, oranges, refined sugar, dairy produce, flour, and rice were all reintroduced without any objective ill effect, but green vegetables provoked a clinical and biochemical relapse within one week of introduction. Remissionwas rapidly achieved by switching back to the enteral feed but reintroduction of the low residue diet that had been previously tolerated produced a brisk relapse. Clinical and biochemical remission was again achieved by a return to the enteral feed but a relapse again occurred with reintroduction of the low residue diet. These studies confirm the therapeutic effect of enteral feeding in Crohn's disease. This effect does not seem to be due to avoidance of whole protein, but the very low residue of chemically defined enteral feeds may be important, particularly in patients with intestinal strictures.There is increasing evidence that a good therapeutic response can be achieved in Crohn's disease by dietary treatment alone. Initially this was thought to be achievable only by complete 'bowel rest,' which was accomplished by intravenous feeding and avoidance of enteral food.'2 Since then it has been shown that bowel rest is not essential34 and that equally good results can be achieved by enteral feeding using a carefully defined 'elemental' feed consisting mainly of amino acids and carbohydrate." It is not clear how enteral feeds achieve their therapeutic effect. Although amino acid based feeds are presumably hypoallergenic, in a recent trial good results were reported with a feed containing whole egg albumin.'0 Furthermore, the chronic inflammation that typifies the inflammatory response in ...