The high nitrogen demands imposed by severe trauma may quickly render the injured patient malnourished. Nutritional assessment of these patients is confused by tissue damage, shock, blood loss, operation, and anesthesia. Consequently, standard nutritional markers do not correlate well with immunocompetence and postoperative morbidity. For this reason we devised an abdominal trauma index (ATI) based on the anatomical severity of injury. The ATI is calculated by assigning a risk factor (1-5) to each organ injured and then multiplying this by a severity-of-injury estimate (1-5). The sum of the individual organ scores comprises the final ATI. The incidence of postlaparotomy complications is low (5%) with an ATI less than 15, intermediate (15%) with 15-25, and high (50%) with greater than 25. Having identified the high-risk trauma patient, we initiated a prospective randomized study to assess the cost-benefit of early nutritional support. Patients with an ATI greater than 15 were allocated to a control group (no supplemental nutrition during first five postoperative days) or enteral-fed group. The enteral group had a needle catheter jejunostomy (NCJ) placed at laparotomy. The constant infusion of an elemental diet (Vivonex HN) was begun at 18 hours postoperatively and advanced to 3,000 cc/day within 72 hours. To date 26 patients (14 control, 12 enteral) have been entered in this study. At one week, nitrogen balance in the control group (−12.9 to -11.1 g/day) continues to be negative compared to a positive trend (−12.2 to +3.3 g/day) in the fed group. In control patients serum albumin (3.54 +/−0.16 to 3.19 +/−0.15 g%) and transferrin (227 +/−11 to 204 +/−10 mg%) decrease while in the enteral patients albumin (3.27 +/−0.11 to 3.34 +/−0.15 g%) and transferrin (229 +/−10 to 234 +/−12 mg%) remain stable. Although the incidence of overall morbidity is similar, septic complications occurred in 29% (4/14) of the control group compared to none in the enteral group. Our experience suggests the following: (1) Anatomical severity of injury is a better predictor of postinjury septic morbidity than standard nutritional markers; (2) immediate postoperative feeding by NCJ is safe and feasible; and (3) early nutritional support decreases the incidence of septic complications in the severely injured patient.