We developed a fixed-volume porcine hemorrhage model that simulates the rapid exsanguination of combat or civilian trauma victims. In this study we compared the ability of colloid resuscitation solutions to prevent death after an otherwise lethal hemorrhage in 100 swine. The shed blood was replaced in a 1:1 ratio with either autologous whole blood (WB), untyped swine fresh frozen plasma (FFP), typed FFP, 5% human serum albumin (ALB), or normal saline (NS). Survival rate analysis indicated that WB was significantly better than FFP (untyped), ALB, or NS but not better than typed FFP. The 24-hour survival rates were: WB = 90%, typed FFP = 79%, untyped FFP = 56%, ALB = 57%, and NS = 25%. All deaths in the untyped FFP group suddenly occurred during or within 15 minutes after treatment in a recovering animal. Deaths in the ALB group steadily occurred for up to 2 1/2 hours after treatment. Analysis of hemodynamic, arterial blood gas, and acid-base data indicated that WB and FFP provided a better acid-buffering capacity in surviving animals than NS or ALB. We conclude that compatible FFP is a better resuscitation agent than ALB after an otherwise fatal hemorrhage because FFP is a better acid buffer.