We continuously recorded systemic and renal haemodynamic changes, and arterial, renal venous and urinary concentrations of thromboxane B2, 6-keto-prostaglandin F1 alpha and prostaglandin E2, and determined their relationship to renal function in an ovine model of progressive hyperdynamic sepsis. 2. Nine chronically instrumented unanaesthetized sheep were given a continuous intravenous infusion of Escherichia coli endotoxin (20 ng min-1 kg-1) for 3 days. 3. Within the first 12 h of infusion, endotoxin induced a major hypotensive septic syndrome, including a persistent 30% reduction in mean arterial pressure, a 50% decrease in systemic vascular resistance and a 50% increase in mean pulmonary artery pressure, associated with severe lactacidaemia. 4. Renal blood flow decreased by 40%, and creatinine clearance, urine flow, and fractional sodium excretion decreased by more than 75%, of baseline values. After 12 h of endotoxin infusion, cardiac output increased two-fold and renal blood flow recovered to baseline values, whereas creatinine clearance remained depressed. Four sheep died between 13 and 22 h of endotoxaemia; these animals (allocated to group 1) presented a significantly and persistently more reduced renal blood flow (-23%) and creatinine clearance (-77%) after 4 h than the remaining five sheep (allocated to group 2), which survived more than 36 h (-16% and -21%, respectively), whereas systemic and pulmonary haemodynamic and gas exchange data remained similar in both groups. 5. The more pronounced decreases in renal blood flow, creatinine clearance and urine flow in group 1 were associated with higher plasma renin activity and plasma 6-keto-prostaglandin F1 alpha concentrations and a lower fractional urinary excretion of 6-keto-prostaglandin F1 alpha than in group 2, whereas plasma thromboxane B2 concentrations were similarly increased in both groups. Plasma prostaglandin E2 concentrations and urinary excretion were not notably affected by endotoxin infusion in either group. 6. Our results are not in favour of a significant renal production of any of these three prostanoids during endotoxaemia. In both groups, values of creatinine clearance were linearly correlated with simultaneous mean arterial pressure values after starting endotoxin infusion (group 1: creatinine clearance = 1.99 x mean arterial pressure--105, r = 0.95; group 2: creatinine clearance = 2.06 x mean arterial pressure--104, r = 0.80).(ABSTRACT TRUNCATED AT 400 WORDS)
