Essential hypertension is associated with endothelial dysfunction, which is caused mainly by the production of oxygen-free radicals that can destroy nitric oxide (NO), and impair its beneficial and protective effects on the vessel wall. In prospective studies, endothelial dysfunction is associated with increased incidence of cardiovascular events. Antihypertensive drugs show contrasting effects in terms of improvement or restoration of endothelial function. Little evidence is available with beta-blockers. Whereas treatment with atenolol has a negative effect in peripheral subcutaneous and muscle microcirculation, insufficient evidence is available to establish whether new compounds such as nebivolol, which activates the L-Arginine--NO pathway, and carvedilol, which has strong antioxidant activity, can improve endothelial function in patients with hypertension. Calcium channel antagonists, particularly the dihydropyridines, can reverse impaired endothelium-dependent vasodilation in different vascular districts, including the subcutaneous, epicardial, renal and forearm circulation. However, conflicting results are found in the brachial artery. In the forearm circulation, nifedipine and lacidipine can improve endothelial dysfunction by restoring NO availability through a mechanism probably related to an antioxidant effect. ACE inhibitors, on the other hand, seem to improve endothelial function in subcutaneous, epicardial, brachial and renal circulation, whereas they are ineffective in potentiating the blunted response to acetylcholine in the forearm of patients with essential hypertension. They can also selectively improve endothelium-dependent vasodilation to bradykinin, an effect not mediated by restoring NO availability but probably related to hyperpolarisation. Recent evidence suggests angiotensin II AT(1)-receptor antagonists can restore endothelium-dependent vasodilation to acetylcholine in subcutaneous microcirculation but not in that of the forearm muscle. Evidence concerning the effect of these drugs on the brachial artery in patients with atherosclerosis is positive. However, treatment with an AT(1)-receptor antagonist can improve basal NO release and decrease the vasoconstrictor effect of endogenous endothelin-1. In conclusion, despite the considerable evidence that impaired endothelium-dependent vasodilation can be improved by appropriate antihypertensive treatment, no clinical data exist demonstrating that the reversal of endothelial dysfunction is associated with a reduction in cardiovascular events. In the near future, large scale clinical trials are required to demonstrate that treatment of endothelial dysfunction can lead to better prognosis in patients with essential hypertension.