OBJECTIVE: To compare the effects of isocaloric, isonitrogenous carbohydrate nutrition vs. lipid-based total parenteral nutrition on respiratory gas exchange and intermediary metabolism in critically ill patients. DESIGN: Prospective, clinical trial. SETTING: Surgical intensive care unit in a major university hospital in Switzerland. PATIENTS: Sixteen patients admitted to the surgical intensive care unit. INTERVENTIONS: Patients were randomized to receive isocaloric isonitrogenous total parenteral nutrition (TPN) containing 75% (TPN-glucose) or 15% (TPN-lipid) glucose over a 5-day period. MEASUREMENTS AND MAIN RESULTS: Indirect glucose metabolism was assessed from plasma carbon-13 (13C)-labeled glucose and 13C-labeled CO2 production during a tracer infusion of uniformly 13C-labeled glucose, and de novo lipogenesis was estimated from the incorporation of 13C into palmitate-very low density lipoproteins (VLDL) during a tracer infusion of 1-(13)C acetate. Compared with TPN-lipid, TPN-glucose increased plasma glucose more (by 26% vs. 7%, p < .05), increased insulin more (by 284% vs. 40%, p < .01), and increased total CO2 more (by 15% vs. 0%, p < .01). Both nutrient mixtures failed to inhibit endogenous glucose production and net protein oxidation, suggesting absence of suppression of gluconeogenesis. Fractional de novo lipogenesis was markedly increased by TPN-glucose to 17.4% vs. 3.3% with TPN lipids. CONCLUSIONS: The rate of glucose administration commonly used during TPN of critically ill patients does not suppress endogenous glucose production or net protein loss, but markedly stimulates de novo lipogenesis and CO2 production. Increasing the proportion of fat may be beneficial, provided that lipid emulsion has no adverse effects.