We have developed a conscious septic dog model suitable for in vivo tracer studies. Dogs weighing 10 to 20 kg underwent general anesthesia followed by the insertion of long-term arterial, venous, and portal cannulas and the formation of a long-term tracheostomy. After 7 to 10 days of convalescence, the animals were fed in the morning and 4 hours later 10(10) live Escherichia coli organisms were infused intra-arterially over approximately 30 minutes. One hour later a second dose of 5 X 10(9) bacteria was given, again over 30 minutes. Resuscitation was provided by infusion of 1000 ml of lactated Ringer solution over 3 hours. Twenty-four hours after the induction of sepsis the animals were hemodynamically stable and suitable for study. Cardiac output was increased from the control value of 185 +/- 35 ml/kg X min to 308 +/- 44 ml/kg X min in the septic animals. Heart rate was increased from 98 +/- 10 to 125 +/- 5 beats/min, and arterial pressure was not significantly altered. We employed indirect calorimetry and primed constant infusions of both radioactive and stable isotopes to assess a variety of metabolic parameters. The metabolic rate was increased approximately 25%, and the energy for this increase was primarily provided by the increased oxidation of both free fatty acids and triglyceride. The release of free fatty acids was approximately three times greater than the control value, and triglyceride synthesis increased 500%. The oxidation rate of free fatty acids and the fatty acids contained in very low density lipoproteins-triglyceride increased 40% and 900%, respectively. Glucose production was maintained at approximately the control value, and the rate of glucose oxidation (as measured with 14C-glucose) was also not significantly altered. The plasma insulin concentration was moderately elevated, and plasma glucagon concentration was five to six times greater than the control value. Plasma catecholamine levels were increased significantly. This model is suitable for the performance of metabolic studies in sepsis. The induction of a hyperdynamic septic state in less than 24 hours avoids the complications of starvation and dehydration frequently seen in the various peritonitis and abscess models. Most importantly, the model is predictable in its time course and reproducibly creates a situation that hormonally, hemodynamically, and metabolically resembles what is commonly seen in humans with sepsis.