Clinical monitoring of cellular metabolism during shock, based largely on traditional metabolic indicators, remains unsatisfactory. The purpose of this study was to compare venous oxygen tension and blood lactate gradients with blood gradients of purine nucleotide degradation products which are derived from tissue ATP catabolism during hypovolemic shock. Sixteen dogs were instrumented to sample arterial and venous blood. Measurements of arteriovenous lactate and PNDP gradients during spontaneous respiration were examined at four tissue sites: gut, kidney, hindlimb, and diaphragm. Hypovolemic shock (mean arterial blood pressure 35 to 40 mm Hg) was induced and maintained for one hour. The above parameters were remeasured at 30 and 60 minutes after induction of shock. Hypoxanthine gradients were greater than that of other PNDP, and so were used as the primary indicator of tissue ATP metabolism. In the hindlimb, the mean AV gradients for hypoxanthine (1 +/- 1 microM) were not significantly greater than baseline, while the lactate gradient (700 +/- 300 microM) rose markedly. In contrast, across the kidney there was a significantly greater AV hypoxanthine gradient (16 +/- 3 microM, p less than 0.002) but no lactate gradient (-400 +/- 200 microM). Both the hypoxanthine and lactate AV gradients were significantly elevated across the diaphragm and gut. Venous PO2 values less than 35 mm Hg predicted an increased hypoxanthine gradient across the kidney, but not across the hindlimb. We conclude that the metabolic response to hypovolemic shock as assessed by PNDP gradients, lactate gradients, and venous PO2 differs among tissues. Although resting muscle such as the hindlimb may be an important source of blood lactate, the viscera and working skeletal muscle (the diaphragm) are major contributors to circulating PNDP.