BACKGROUND: Conversion of glucose into lipid (de novo lipogenesis; DNL) is a possible fate of carbohydrate administered during nutritional support. It cannot be detected by conventional methods such as indirect calorimetry if it does not exceed lipid oxidation. OBJECTIVE: The objective was to evaluate the effects of carbohydrate administered as part of continuous enteral nutrition in critically ill patients. DESIGN: This was a prospective, open study including 25 patients nonconsecutively admitted to a medicosurgical intensive care unit. Glucose metabolism and hepatic DNL were measured in the fasting state or after 3 d of continuous isoenergetic enteral feeding providing 28%, 53%, or 75% carbohydrate. RESULTS: DNL increased with increasing carbohydrate intake (f1.gif" BORDER="0"> +/- SEM: 7.5 +/- 1.2% with 28% carbohydrate, 9.2 +/- 1.5% with 53% carbohydrate, and 19.4 +/- 3.8% with 75% carbohydrate) and was nearly zero in a group of patients who had fasted for an average of 28 h (1.0 +/- 0.2%). In multiple regression analysis, DNL was correlated with carbohydrate intake, but not with body weight or plasma insulin concentrations. Endogenous glucose production, assessed with a dual-isotope technique, was not significantly different between the 3 groups of patients (13.7-15.3 micromol * kg(-1) * min(-1)), indicating impaired suppression by carbohydrate feeding. Gluconeogenesis was measured with [(13)C]bicarbonate, and increased as the carbohydrate intake increased (from 2.1 +/- 0.5 micromol * kg(-1) * min(-1) with 28% carbohydrate intake to 3.7 +/- 0.3 micromol * kg(-1) * min(-1) with 75% carbohydrate intake, P: < 0. 05). CONCLUSION: Carbohydrate feeding fails to suppress endogenous glucose production and gluconeogenesis, but stimulates DNL in critically ill patients.