1. Antecubital venous plasma glucose and insulin concentrations after ingestion of 75 g of glucose were higher in six normal subjects when studied at an ambient temperature of 33°C than at an ambient temperature of 23°C; the mean area under the glucose-time curve increased from 833 at 23°C to 990 mmol min 1 −1 at 33°C, that for insulin from 5300 to 7900 m-units min 1 −1 . 2. Core temperature was elevated by 0.5°C at 33°C ambient, although there was no marked stress response as judged by plasma levels of catecholamines, cortisol and growth hormone; at 2 h after glucose ingestion, mean noradrenaline levels were lower at 33°C than at 23°C (1.1 at 33°C vs 1.8 nmol/l at 23°C), adrenaline slightly higher (0.18 at 33°C vs 0.09 nmol/l at 23°C), cortisol and growth hormone unchanged. 3. Forearm blood flow was markedly elevated at 33°C ambient (mean total flow 9.1 at 33°C vs 1.8 ml min −1 100 ml −1 at 23°C), as were antecubital venous partial pressure of oxygen (mean 10.1 at 33°C vs 5.6 kPa at 23°C) and oxygen saturation (mean 92% at 33°C vs 70% at 23°C). There was a positive correlation between oxygen saturation and area under the glucose tolerance curve. 4. In separate experiments, arterialized glucose concentrations were measured after glucose ingestion at 23°C ambient. Antecubital venous glucose concentrations at 33°C ambient were intermediate between venous and arterialized concentrations at 23°C ambient. 5. It is concluded that the apparent impairment of glucose tolerance at elevated ambient temperature is caused at least in part by increased arterialization of the antecubital venous blood. This potential variable must be considered in research or diagnostic procedures based on the use of venous blood.

1. Plasma concentrations of glucose, lactate, amino acids, non-esterified fatty acids, glycerol, ketone bodies, insulin and cortisol were measured in 43 elderly patients with hypothermia. In 15 of these patients forearm arteriovenous differences were also measured. Core temperatures ranged from 25.9 to 35.5°C. 2. The metabolic state was of mobilization of glycogen and triacylglycerol stores, with high plasma concentrations of lactate and lipid metabolites. The plasma concentration of glucose was raised in those with hypothermia of a short duration (<6 h). In other patients it was low in those with core temperatures around 30°C, but below this temperature it was variable and often high. Concentrations of other metabolites or hormones were not related to core temperature. 3. Plasma concentrations of cortisol were high and positively correlated with those of lactate and glycerol, suggesting active involvement in stimulation of muscle glycogenolysis and of lipolysis. 4. Plasma concentrations of insulin ranged from very low to very high and appeared to depend on the concentrations of both glucose and alanine. 5. Arteriovenous differences were generally small. There was peripheral release of lactate and of amino acids but no overall peripheral uptake of glucose. In nine out of 15 patients there was a significant peripheral release of glucose. 6. No differences in metabolism were observed between patients where the hypothermia appeared accidental and those with an obvious precipitating illness, despite a significantly lower mortality in the former group. 7. It was concluded that therapy should primarily involve rewarming of patients by physical means, without metabolic intervention.