Several controlled, experimental, hypotensive models of hemorrhagic shock have evaluated the effects of timing, rate, and types of fluid replacement. In a near-fatal experimental model we evaluated the hemodynamic and metabolic effects of two types of solutions for fluid resuscitation. In this study, 30 young Large-White pigs were randomly assigned to three groups: Group I (control, n= 10), not bled; Group II (hydroxyethyl starch, HES, n = 10), submitted to controlled hemorrhage to a mean arterial blood pressure (MAP) of 30 mmHg and blood lactate >10 mM/L, at which time resuscitation was initiated with 7 mL/kg of HES 130/0.4 6% followed by 33 mL/kg of lactated Ringer's solution (LR) and retransfusion; Group III (LR, n = 10), submitted to controlled hemorrhage to a MAP of 30 mmHg and blood lactate >10 mM/L, at which time resuscitation was initiated with 40 mL/kg of LR followed by retransfusion. The resuscitation with HES 130/0.4 proved to be superior to LR, expressed by hemodynamic and perfusion variables. Despite improvement in tissue perfusion, MAP did not totally return to baseline values. In conclusion, early colloid infusion resulted in prompt recovery of tissue perfusion when compared with infusion with an equal volume of crystalloid.