Controversial studies have been published concerning the role of nitric oxide (NO) release (beneficial or deleterious) during sepsis. Severe hypotension has been treated by NO inhibitors in humans, but animal studies described an increased mortality rate with this treatment. We hypothesized that an NO donor might be beneficial in maintaining liver flow during endotoxemia. To answer that question, mean arterial pressure (MAP), aortic, hepatic artery, and portal vein blood flow velocities (AoV, HAV, and PVV) (Doppler technique) were measured after endotoxin injection (Escherichia coli, Salmonella minnesota, and Salmonella enteritidis, 400 micrograms each, intravenously) in anesthetized and mechanically ventilated rabbits. Fifteen animals were treated with saline solution (10 mL/hr) or linsidomine perfusion (2 mg over 3 hours, 10 mL/hr). Saline-treated animals experienced a hypodynamic shock with a decrease in MAP, AoV, and PVV. In contrast, HAV increased without fully compensating the PVV decrease. In linsidomine-perfused rabbits, AoV and PVV remained at control level, and HAV increased without any further effect on MAP. Serum lactate levels increased in the saline-treated group and did not change in linsidomine-treated animals. These findings show that at the early phase of an endotoxin shock, and in the absence of intense fluid resuscitation, linsidomine perfusion is beneficial in maintaining systemic and hepatic perfusion while preventing lactic acidosis. These data suggest that, in the early phase of endotoxemia, NO is insufficiently released to allow adequate liver perfusion.