Epinephrine is widely used for treatment of life-threatening hypotension, although new vasopressor drugs may merit evaluation. The purpose of this study was to determine the effects of vasopressin vs. epinephrine vs. saline placebo on hemodynamic variables, regional blood flow, and short-term survival in an animal model of uncontrolled hemorrhagic shock and delayed fluid resuscitation. Prospective, randomized, laboratory investigation that used a porcine model for measurement of hemodynamic variables and regional abdominal organ blood flow. University hospital laboratory. A total of 21 pigs weighing 32 +/- 3 kg. The anesthetized pigs were subjected to a penetrating liver injury, which resulted in a mean +/- sem loss of 40% +/- 5% of estimated whole blood volume within 30 mins and mean arterial pressures of <20 mm Hg. When heart rate declined progressively, pigs randomly received a bolus dose and continuous infusion of either vasopressin (0.4 units/kg and 0.04 units.kg-1.min-1, n = 7), or epinephrine (45 microg/kg and 5 microg.kg(-1).min(-1), n = 7), or an equal volume of saline placebo (n = 7), respectively. At 30 mins after drug administration, all surviving animals were fluid resuscitated while bleeding was surgically controlled. Mean +/- sem arterial blood pressure at 2.5 and 10 mins was significantly (p <.001) higher after vasopressin vs. epinephrine vs. saline placebo (82 +/- 14 vs. 23 +/- 4 vs. 11 +/- 3 mm Hg, and 42 +/- 4 vs. 10 +/- 5 vs. 6 +/- 3 mm Hg, respectively). Although portal vein blood flow was temporarily impaired by vasopressin, it was subsequently restored and significantly (p <.01) higher when compared with epinephrine or saline placebo (9 +/- 5 vs. 121 +/- 3 vs. 54 +/- 22 mL/min and 150 +/- 20 vs. 31 +/- 17 vs. 0 +/- 0 mL/min, respectively). Hepatic and renal artery blood flow was significantly higher throughout the study in the vasopressin group; however, no further bleeding was observed. Despite a second bolus dose, all epinephrine- and saline placebo-treated animals died within 15 mins after drug administration. By contrast, seven of seven vasopressin-treated animals survived until fluid replacement, and 60 mins thereafter, without further vasopressor therapy (p <.01). Moreover, blood flow to liver, gut, and kidney returned to normal values in the postshock phase. Vasopressin, but not epinephrine or saline placebo, improved short-term survival in a porcine model of uncontrolled hemorrhagic shock after liver injury when surgical intervention and fluid replacement was delayed.