Although nutritional support is vital to treatment of severe sepsis, the septic patient does not respond normally to glucose infusion. We have used the hyperglycemic glucose clamp technique to investigate the initial hormonal and metabolic responses of the septic patient to glucose under controlled conditions. The plasma glucose concentration was raised to and maintained at 12 mmol/liter for 2 hr in 12 septic patients and 11 normal controls. Glucose utilization, assessed from the amount infused, was significantly depressed in the patients, despite similar plasma insulin concentrations in the two groups. Forearm glucose uptake was similarly impaired. Despite very similar plasma free fatty acid concentrations in the two groups, which were suppressed equally by the glucose infusion, whole-body fat oxidation was elevated in the patients compared with the controls, and suppressed to a lesser extent in response to glucose. Glycerol and ketone body concentrations were elevated in the patients in keeping with a picture of accelerated release, clearance, and oxidation of fatty acids. Plasma cortisol, epinephrine, and norepinephrine concentrations were elevated in the septic patients in a severity-related manner, but not to high levels compared with experimental work. Norepinephrine showed no response to the glucose infusion in either group. Plasma glucagon concentrations were not significantly elevated in the septic patients. We conclude that the hyperglycemic glucose clamp provides a useful model for studying glucose intolerance in sepsis. Impaired glucose utilization in septic patients is associated with increased fat oxidation, although the hormonal basis for these changes is still unclear.

1. The plasma concentrations of glucose, lactate, amino acids, non-esterified fatty acids, glycerof, ketone bodies, ethanol, cortisol and insulin were measured in patients withii a few hours of injury and before treatment.' The severity of the injuries was assessed by the Injury Severity Score (ISS) method.2. Plasma lactate and glucose concentrations both rose significantly with increasing ISS.3. The concentrations of non-esterified fatty acids and glycerol were greater after moderate (ISS 7-12) than after minor (ISS 1-6) injuries. The glycerol concentrations were no higher and the non-esteritied fatty acid concentrations were lower after severe (ISS > 12) than after moderate injuries. The concentrations of total ketone bodies tended to follow those of non-esterified fatty acids and there was a highly significant correlation between them. 4. The total concentration of amino acids was not affected by the severity of injury and there were no systematic changes in the concentrations of individual ones. 5. Plasma insulin concentrations were very variable and not related to severity. A weak correlation with the plasma glucose concentration seen after minor and moderate injuries was lost in the severely injured. Correspondence: Professor H. B. Stoner, MRC Trauma Unit, Stopford Building, University of Manchester, Oxford Road, Manchester M13 9PT, U.K.6. The plasma cortisd concentration was positively related to ISS up to ISS 12 but negativeIy so in the severely injured.7. Factors such as age, sex and time after last meal were investigated. The most important factor modifying the response was intake of ethanol, which reduced the plasma concentrations of glucose, non-esterilied fatty acids and alanine and raised that of lactate as well as the hydroxybutyratel/LacetoacetateI ratio,

1. The equations derived by Heath (1968) were applied to data from experiments on rats in four metabolic states: fed, post-absorptive, starved and 2hr. after an eventually lethal injury. The data used were: (a) The fractions of label injected as C1-, C2- and C3-pyruvate (where the prefix indicates the position of labelling) that are incorporated into carbon dioxide and glucose in post-absorptive and injured rats (yields). Yields could be corrected to yields on label taken up by the liver. (b) The (C5-label in glutamate)/(total label in glutamate) ratio in the liver after C2-pyruvate in rats in all four states. (c) The distribution of label within glutamate after C2-pyruvate or C2-alanine in the livers of fed, post-absorptive and starved rats. (d) The distribution of label within glucose after C2-lactate or C2-pyruvate in starved rats. (e) The relative specific radioactivities of pyruvate, aspartate, glutamate and (in two states only) of glucose 6-phosphate after injection of [U-(14)C]glucose into rats in all four states. These data were previously published, except those after (e) and some after (b) above, which are given in this paper. 2. In addition the concentrations of pyruvate, citrate, glutamate and aspartate in the livers of post-absorptive and injured rats were found. Injury decreased glutamate and citrate concentrations and to a smaller extent aspartate and pyruvate concentrations. 3. Non-steady-state theory showed that most of the data could be used without serious error in steady-state theory. Steady-state theory correlated all but one observation (the relative yields of (14)CO(2) from C2- and C3-pyruvate) listed after (a)-(e) above within the experimental errors, and gave rough estimates of the rates of pyruvate carboxylation, conversion of pyruvate and fat into acetyl-CoA and utilization of glutamate. The main conclusions were: (a) symmetrization of label in oxaloacetate both in the mitochondrion and in the cytoplasm was far from complete, because oxaloacetate did not equilibrate with fumarate in either. From this and other findings it was deduced: (b) that malate or fumarate or both left the mitochondrion, and not oxaloacetate; (c) that there was a loss from the mitochondrion of a fraction of the malate or fumarate or both formed from succinate, and (d) the resulting deficiency of oxaloacetate for the perpetuation of the tricarboxylic acid cycle was made up from pyruvate in fed and post-absorptive rats, but (e) in the starved rat could only be made up by utilization of glutamate. (f) In the fed rat the tricarboxylic acid cycle ran mostly on pyruvate, but in the post-absorptive and starved rat mostly on fat. (g) In the injured rat the tricarboxylic acid cycle was slowed, label in oxaloacetate was completely symmetrized (cf. conclusion a), and the tricarboxylic acid cycle utilized glutamate. (h) The conclusions were not invalidated by isotopic exchange, i.e. flux of label without net flux of compound, nor by interaction with lipogenic processes. (i) In the kidneys interaction between the tricarboxylic...