This two-part meta-analysis combined data from eight prospective randomized trials designed to compare the nutritional efficacy of early enteral (TEN) and parenteral (TPN) nutrition in high-risk surgical patients. The combined data gave sufficient patient numbers (TEN, n = 118; TPN, n = 112) to adequately address whether route of substrate delivery affected septic complication incidence. Phase I (dropouts excluded) meta-analysis confirmed data homogeneity across study sites, that TEN and TPN groups were comparable, and that significantly fewer TEN patients experienced septic complications (TEN, 18%; TPN, 35%; p = 0.01). Phase II meta-analysis, an intent-to-treat analysis (dropouts included), confirmed that fewer TEN patients developed septic complications. Further breakdown by patient type showed that all trauma and blunt trauma subgroups had the most significant reduction in septic complications when fed enterally. In conclusion, this meta-analysis attests to the feasibility of early postoperative TEN in high-risk surgical patients and that these patients have reduced septic morbidity rates compared with those administered TPN.