Endotoxin given intravenously is known to cause plasma leakage and subsequent loss of circulating plasma volume. Hypertonic saline resuscitation has been successfully applied in hemorrhagic and traumatic shock, but its application for the treatment or prevention of septic or endotoxin shock is less well studied. Our aim was to investigate the effects of endotoxin on plasma leakage in hamsters when administered in two different ways: applied locally to the hamster cheek pouch microcirculation or systemically by i.v. injection. The cheek pouch was studied by intravital microscopy using FITC-labeled dextran as a tracer of plasma leakage. Escherichia coli lipopolysaccharide (LPS) was continuously added into the superfusion buffer of the cheek pouch preparation during 120 min in two control groups (each n = 6) and two further groups (each n = 6) treated with either hypertonic saline (HS) or hypertonic saline and dextran (HSD). Treatment was given as an i.v. injection 0.35 mL NaCl 7.5%/100 g b.w. during 4 min starting 15 min prior to the start of endotoxin application. Endotoxin caused a reversible increase in the number of postcapillary venular leaks with a maximal response at 70 min after start of endotoxin application. The maximal responses were reduced to 36% in the HS-treated and to 37% in the HSD-treated group in comparison to what was seen in the control groups. In the second part of the study endotoxin was given i.v. 0.3 mg/kg to anesthetized hamsters (n = 41) and arterial blood samples were collected at 0, 60, 120, and 180 min after endotoxin injection for measurement of hematocrit and plasma FITC-dextran concentration. Hamsters were divided into seven groups: untreated control group (n = 6); HSC control group given only an i.v. injection of hypertonic saline (n = 6); LPS group given endotoxin 0.3 mg/kg during 1 min (n = 9); HSp group given hypertonic saline (NaCl 7.5%) 10 min prior to i.v. endotoxin (n = 6); HSa group given hypertonic saline 10 min after i.v. endotoxin (n = 6); HSD group given hypertonic saline with dextran 40, 10 min prior to i.v. endotoxin (n = 6); HSD control group given only i.v. hypertonic saline + dextran and no endotoxin (n = 2). Injection of endotoxin caused a significant increase in hematocrit, which was counteracted by hypertonic saline treatment, with or without dextran, probably due to reduced extravasation of plasma in postcapillary venules.