The vascular endothelium achieved a critical place in the understanding of vascular physiology and pathophysiology, after the discovery of the production of prostacyclin by endothelial cells, followed by the recognition that substances like acetylcholine, assumed to be direct vasodilators, could only trigger dilation in the presence of an intact endothelium. The endothelium-derived relaxing factor (EDRF) behaves as an endogenous nitrovasodilator and causes vasodilatation through stimulation of guanylyl cyclase and cellular accumulation of cyclic GMP. Subsequently, it was demonstrated that the EDRF is nitric oxide (NO), produced through the metabolism of the aminoacid L-arginine by the nitric oxide synthases (NOS). Three isoforms of this enzyme were discovered and cloned: a constitutive neuronal isoform (nNOS); an inducible isoform (iNOS), ubiquitous in cells stimulated by certain cytokines; and an endothelial isoform (eNOS). The importance of the different isoforms is well demonstrated in animal models; more recently, human studies unveiled the importance of these enzymes. The endothelium produces other vasodilators besides NO and prostacyclin; furthermore, it produces several vasoconstrictors. There is a delicate balance between these factors, which can be disturbed: several well known cardiovascular aggressors, like arterial hypertension, diabetes, smoking, dyslipidemia or renal insufficiency, can alter several invasive or non-invasive tests of endothelial function. The fact that an intervention on these factors may reverse endothelial dysfunction as measured by these tests, raises hope that they may be surrogate markers of global cardiovascular risk. If correlation of these tests with clinical outcomes proves to be robust, they may become extensively used in clinical practise.