The aim of this study was to confirm that microvascular perfusion was abnormal during the early phases of normotensive sepsis and to determine whether these changes were due to the development of tissue edema. Skeletal muscle red blood cell (RBC) flow was studied in rats made septic by cecal ligation and perforation (CLP). After anesthesia with halothane, arterial and venous cannulae were inserted and, in the treatment group, a CLP performed. At 6, 24, and 48 h after entry into the study, the incidence of microcirculatory absence of flow in the extensor digitorum longus muscle (EDL) was examined with intravital microscopy. The number of capillaries containing RBCs were counted over a 60-s interval, and the flow status of each capillary was recorded. A significant increase in the number of stopped-flow capillaries was observed in the CLP group (p < 0.01) as compared with time-matched controls. In both groups the number of capillaries with stopped flow was greater than in naive animals. The severity of absence of flow was negatively correlated with the systemic hemoglobin concentration. These changes were not associated with an increase in tissue wet/dry weight ratio or albumin flux. This study shows that sepsis was associated with increased RBC flow heterogeneity. These changes, which occur within 24 h of the septic insult, are a persistent feature of the evolving septic process in the absence of tissue edema. These observations support the view that extrinsic compression of the microcirculation by tissue edema is not the primary cause of alterations in microcirculatory flow in sepsis.

In animal models of endotoxemia, sepsis is associated with the accumulation of leukocytes and altered microvascular perfusion. In order to test the hypothesis that bacterial sepsis upregulates leukocyte-endothelial adhesion, we used intravital microscopy to examine the flow behavior of leukocytes in the postcapillary venules (PCV) of rats made septic by cecal ligation and perforation (CLP). Animals were randomized to CLP or sham study groups and studied 6 h, 24 h, or 48 h later. In postcapillary venules of the extensor digitorum longus muscle, we found that: (1) over the course of the study, leukocyte adhesion and extravasation increased in both experimental groups (analysis of variance [ANOVA], significant time effect: adhesion, p < 0.001; extravasation, p < 0.05); (2) leukocyte adhesion was decreased by CLP treatment (ANOVA, sepsis effect, p = 0.05), particularly after 24 to 48 h of sepsis (ANOVA, sepsis x time interaction, p < 0.05); and (3) the reduction in leukocyte adhesion in CLP animals was associated with a decrease in leukocyte extravasation (ANOVA, sepsis effect, p < 0.01). After correction for the reduction in systemic leukocyte count associated with CLP, the effect of sepsis on leukocyte adhesion and extravasation no longer reached statistical significance. These findings suggest that chronic (6 to 48 h) bacterial sepsis does not upregulate leukocyte adhesion in a manner similar to that seen in models of acute endotoxemia. These data suggest that the increased microcirculatory flow heterogeneity seen in this and other models of bacterial sepsis may not be explained by leukocyte entrapment in postcapillary venules.