Nitric oxide (NO) is one of many vasoactive substances released, from a variety of cells, under conditions of endotoxaemia and sepsis. Under physiological conditions it is produced by two constitutive calcium-dependent enzymes (nitric oxide synthase; NOS) in neurones (nNOS) and endothelial cells (eNOS) and has functions ranging from neurotransmission and vasodilatation to inhibition of platelet adhesion and aggregation. Following bacterial infection, especially with Gram-negative organisms, its formation from L-arginine is enhanced due to the cytokine-mediated induction of a NOS enzyme (iNOS) in cells (e.g. cardiac myocytes, vascular smooth muscle) that do not normally have the ability to synthesize NO. The result of this excessive NO production is enhanced bacterial lysis by activated macrophages, vasoplegia and myocardial depression. These cardiovascular effects can be alleviated by inhibitors of the L-arginine NO pathway, which results in elevated perfusion pressure, restored responsiveness to sympathetic nerve stimulation and to exogenous catecholamines, and to enhanced (endothelin-dependent) myocardial contractility. In patients in shock this approach also leads to detrimental effects (increased systemic vascular resistance, elevated pulmonary artery pressure, reduced cardiac output and oxygen delivery, increased platelet accumulation) and survival is not improved. Because some of these detrimental effects are due to inhibition of eNOS, attempts have been made to examine the effects of substances with a higher selectivity for the induced form of the enzyme. In experimental animals, one of these (L-canavanine) protects endothelial cells from damage, increases survival time and restores vascular responsiveness without increasing blood pressure or peripheral vascular resistance. However, whether even this approach will be of benefit to patients with sepsis remains in doubt since studies in iNOS knock-out mice do not support the concept that eliminating this particular source of NO improves ultimate survival.