A randomized controlled prospective clinical trial has been undertaken to examine the efficacy of the technique of early postoperative feeding using a fine bore catheter jejunostomy. Fifty patients undergoing surgery for gastrointestinal malignancy were randomly allocated into treatment and control groups. A low residue liquidized diet (Isocal) was administered to the patients in the treatment group. Control patients received routine intravenous therapy. Nutritional parameters (serum albumin, serum transferrin, serum prealbumin, weight, body fat and fat free mass) were measured pre-operatively and on the tenth postoperative day. Postoperative surgical complications were similar in both groups. There were 20 catheter complications and one death directly attributable to the jejunal catheter feeding. Postoperative stay was significantly longer (P less than 0.01) in the treatment group patients. Evaluation of the nutritional parameters showed no advantage for either the treatment group or a selected complication-free, 'successful treatment', subgroup. It is concluded that no significant clinical or nutritional advantage for jejunal catheter feeding has been demonstrated and because of the related complications, its routine use cannot be recommended.