Immunonutrition in the critically ill: A systematic review of clinical outcome

Objective: To perform a meta-analysis addressing whether enteral nutrition with immune-enhancing feeds benefits critically ill patients after trauma, sepsis, or major surgery. Data Sources: Studies were identified by MEDLINE search (1967 to January 19 98) for original articles in English using the search terms "human," "enter al nutrition," "arginine," "nucleotides," "omega-3 fatty acids," "immunonut rition," "IMPACT," and "Immun-Aid." Additionally, the authors of the studie s and the manufacturers of the feeds were contacted for additional informat ion. Access to original databases was obtained for the three largest studie s. Study Selection: Fifteen randomized controlled trials comparing patients re ceiving standard enteral nutrition with patients receiving a commercially a vailable immune-enhancing feed with arginine with or without glutamine, nuc leotides, and omega-3 fatty acids were identified by two independent review ers (Dr. Beale and Dr. Bryg). Data Extraction: Descriptive and outcome data were extracted independently from the papers by the same two reviewers, one of whom (Dr. Bryg) analyzed the original databases. Three studies were excluded from analysis, leaving 12 studies containing 1,557 subjects, 1,482 of whom were analyzed. Main out come measures were mortality, infection, ventilator days, intensive care un it stay, hospital stay, diarrhea days, calorie intake, and nitrogen intake. The meta-analysis was performed on an intent-to-treat basis. Data Synthesis: There was no effect of immunonutrition on mortality (relati ve risk = 1.05, confidence interval [CI] = 0.78, 1.41; p = .76). There were significant reductions in infection rate (relative risk = 0.67, CI = 0.50, 0.89; p = .006), ventilator days (2.6 days, CI = 0.1, 5.1; p = .04), and h ospital length of stay (2.9 days, CI = 1.4, 4.4; p = .0002) in the immunonu trition group. Conclusions: The benefits of enteral immunonutrition were most pronounced i n surgical patients, although they were present in all groups. The reductio n in hospital length of stay and infections has resource implications.