Natural vasopressins have been used, with varying success, in attempts to stop bleeding from esophageal varices for over two decades. Reasons for lack of success include (a) failure to induce a sufficiently prolonged and constant vasoconstrictor effect at the bleeding site, (b) dangerous side-effects, and (c) release of plasminogen activator induced by the peptides which can lyse any clot as it forms.IN THE LAST DECADE ANALOGS OF VASOPRESSIN HAVE BEEN DEVELOPED WITH A PROLONGED ACTION, USING TWO SEPARATE PRINCIPLES OF CHEMICAL MODIFICATION: (1) hormonogens, and (2) blockage of sites of inactivating enzymatic cleavage (in particular "carba" analogs without a disulfide bridge). These two categories of analog are compared here: the carba analogs have the advantages of high potency (higher than the parent hormone) with prolongation, but are also very active on the plasminogen activator release system. The hormonogens combine prolongation with low potency, but have lost not only a releasing action on plasminogen activator, but also, by virtue of altered release kinetics, have effectively lost cardiovascular toxicity.Mechanisms of analog action and receptor interaction are presented, along with initial clinical experiences.