Metabolic support is an integral component of surgical critical care. Although prompt restoration of oxygen availability is clearly essential, the timing, composition, and route of nutritional support may also be decisive factors. The ensuing discussion will focus on: (a) timing of substrate delivery and (b) route of administration based on our clinical investigation over the past decade. The acutely injured patient was selected as a model of ICU hypermetabolism because of relative homogeneity with respect to age, comorbid factors, and stress level. Our first study hypothesis was that early nutritional support would improve outcome in the severely injured, but previously well-nourished patient. During an 18-month period, all patients undergoing laparotomy with a abdominal trauma index (ATI) greater than 15 were randomized to a control or total enteral nutrition (TEN) group. The control patients were given total parenteral nutrition (TPN) after POD 5, whereas the TEN cohort had a needle catheter jejunostomy (NCJ) inserted at laparotomy and received an elemental diet within 12 hours. The control (n = 31) and TEN (n = 32) groups were otherwise comparable with respect to risk stratification. The TEN patients, of course, shared improved nitrogen balance (p less than 0.001), but also had significantly (p less than 0.025) less septic morbidity. Nine (29%) of the controls developed major infections, contrasted to three (9%) of the TEN patients. Acknowledging the benefit of early nutrition, the next issue we addressed was the optimal route of substrate delivery; i.e., TEN vs TPN. The hypothesis was that TEN, compared to TPN, would reduce the injury stress response as reflected by the prioritization of hepatic protein synthesis. TEN given via NCJ and a nutritionally matched TPN solution were administered during the same postoperative period. Indeed, the TEN patients (n = 23) had significantly (p less than 0.05) higher constitutive proteins and lower acute-phase proteins, whereas the TPN patients manifested the opposite protein profile as measured by crossed immunoelectrophoresis. In view of these findings, we continued the study to ascertain clinical impact. Ultimately, 75 patients were randomized, providing groups with equivalent risk factors. Eleven (37%) of the TPN patients developed septic complications compared to five (17%) of the TEN group, and the incidence of major infection was six (20%) following TPN vs one (3%) with TEN. Thus, immediate TEN provided an additional clinical benefit compared to early TPN in these high-risk surgical patients.