Little is known about how physiological concentrations of glucocorticoid relate to cardiac metabolism in vivo. Healthy conscious dogs with catheters implanted for blood sampling and glucose infusion were studied. The range of blood glucose values produced by glucose infusion was 3700 to 74,400 mumol/liter. Arterial glucose concentration (Ca) was not significantly correlated with the arterial-coronary sinus difference in concentration of glucose, Ca-cs glucose (N = 50, r = 0.23). However, at or above glucose infusion rates of 2120 mumol/min, significant increases in cardiac glucose extractions were seen. The range of plasma cortisol values was 13 to 438 nmol/liter. Cortisol immunoreactivity in arterial plasma (Ia cortisol) was significantly and negatively correlated with Ca-cs glucose (N = 50, r = -0.37, P < 0.01). Nonparametric analysis confirmed this association (Spearman rank correlation coefficient, rs = -0.44, P < 0.01). The heart took up and released cortisol in relation to Ia cortisol (N = 50, r = 0.75, P < 0.001; rs = 0.58, P < 0.001). The range of Ia-cs cortisol was -31 to 138 nmol/liter (mean +/- SEM, 12 +/- 4, P < 0.01). The Ca-cs glucose was negatively correlated with Ia-cs cortisol (N = 50, r = -0.43, P < 0.01; rs = 0.50, P < 0.001). Thus, higher plasma cortisol immunoreactivity may lead to greater myocardial cortisol extraction and suppression of myocardial glucose extraction in vivo. At the same time arterial insulin immunoreactivity had a significant positive relationship to myocardial glucose extraction (N = 50, r = 0.37, P < 0.05; rs = 0.38, P < 0.01) and arterial plasma free fatty acids a negative relationship (N = 50, rs = -0.30, P < 0.05).