Objective: To compare recombinant hemoglobin solution (rHb1.1) with colloid/blood (CB) resuscitation in a hemorrhagic shock (HS) model based on oxygen debt (O2D). Methods: Twenty-two anesthetized canines (weight 23.3 +/- 0.2 kg) were bled to an O2D of 99.9 +/- 2.1 mL/kg over 60 minutes (estimated Lethal dose 31%), blood loss 65.9 +/- 2.3% of estimated blood volume, -199 g hemoglobin (Hgb). Prospectively randomized resuscitation done in 20 minutes with 120% of shed blood volume, either colloid 60%/blood 60% shed blood volume (CB), 118 g Hgb, or 120% shed blood volume as 5% rHb1.1, 85 g Hgb, and fall in O2D was quantified over 80 minutes, Six animals died during HS, one after CB resuscitation (32% actual mortality), Blood lactate (L) and base deficit (BEA) were related to O2D. Results: Both lactate (L) and BEA quantified O2D during hemorrhage: L = 0.0671 (O2D) + 1.209; r(2) = 0.90, p < 0.0001; BEA = 0.1313 (O2D) + 1.764; r(2) = 0.90, p < 0.0001, but L was a better indicator than BEA of fall in O2D during resuscitation (L = 0.069 (O2D) + 1.083; r(2) = 0.80), Both groups were followed for 7 days after HS and had normal renal and hepatic function by day 7, However, at equal resuscitation volume, rHb1.1 resuscitation with 41% of Hgh loss produced a more rapid initial fall in O2D than CB at 60% of Hgb loss (p < 0.002), rHb1.1 resuscitation also caused a more complete washout of metabolic acids than CB. Conclusions: Lactate and BEA accurately quantify O2D in HS and resuscitation. rHb1.1 replacement is as good as CB with regard to survival, but leads to a more uniform reperfusion and produces a more complete resolution of ischemic acidosis.