Objective: Review of primate studies of Escherichia coli sepsis and endotoxemia with a reexamination of the rationale for diagnosis and treatment of these multistage disorders.

Setting: Animal research and intensive care units in a university medical school.

Subjects: Cyanocephalus baboons (E. coli) and normal human subjects (endotoxin).

Interventions: Baboon studies: anti-tissue factor, protein C, endothelial protein C receptor, and anti-tumor necrosis factor antibodies, and active site inhibited factor recombinant VIIa and factor Xa.

Results and conclusions: This review concerns the primate microvascular endothelial response to inflammatory and hemostatic stress. Studies of the impact of inflammatory and hemostatic stress on this microvasculature have fallen into four categories. First, studies of pure hemostatic stress using factor Xa phospholipid vesicles showed that blockade of protein C as well as protein C plus tissue plasminogen activator produced a severe but transient consumptive and a lethal thrombotic coagulopathy, respectively. These studies showed that the protein C and fibrinolytic systems can work in tandem to regulate even a severe response if the endothelium is not rendered dysfunctional by metabolic or inflammatory factors. Second, studies of compensated (nonlethal) inflammatory stress using E. coli or endotoxin in baboon and human subjects showed that even under minimal stress in which there is no evidence of overt disseminated intravascular coagulation, injury of the endothelium and activation of neutrophils and hemostatic factors are closely associated. This showed that molecular markers of hemostatic activity could be used to detect microvascular endothelial stress (nonovert disseminated intravascular coagulation) in patients who are compensated but at risk. These studies also showed that the compensated response to inflammatory stress could exhibit two stages, each with its unique inflammatory and hemostatic response signature. The first is driven by vasoactive peptides, cytokines, and thrombin, followed 12 to 14 hrs later by a second stage driven by C-reactive protein/complement complexes, tissue factor, and plasminogen activator inhibitor 1 secondary to oxidative stress after reperfusion. Third, studies of uncompensated (lethal) inflammatory stress using E. coli showed that irreversible thrombosis of the microvasculature was not a link in the lethal chain of events even though inhibition of components of the protein C network (protein C and endothelial protein C receptor) converted compensated responses to sublethal E. coli into uncompensated lethal responses. Fourth, these studies also showed that there were variants of the lethal response ranging from capillary leak and shock to recurrent sustained inflammatory disorders. We believe that each of these variants arises from their sublethal counterparts, depending on underlying or modulating host factors operating at the time of challenge. Such underlying conditions range from preexisting microvascular ischemia, reperfusion, and oxidative stress to alteration or reprogramming of monocyte/macrophage responses (tolerance to hyperresponsiveness). Characterization of these underlying conditions in patients who are at risk should aid in identifying and optimizing management of these variants.