Background: Endothelial dysfunction is presumed to occur after hemorrhagic shock and resuscitation. This study uses a novel large-animal model to evaluate the effects of diverse resuscitation regimens on endothelial function. Methods: Twenty-seven adult domestic pigs (Sus scrofa) were used in this study. Control pigs (n = 3) underwent instrumentation alone. The remaining pigs experienced controlled hemorrhagic shock to a mean arterial blood pressure of 30 ± 5 mm Hg for 45 minutes. Pigs were resuscitated to their baseline mean arterial blood pressure ±5 mm Hg with either shed blood (SB; n = 8), lactated Ringers solution (40 mL/kg) followed by shed blood (LRSB; n = 8), or lactated Ringers solution alone (LR; n = 8). At baseline, 1 and 4 hours after resuscitation, acetylcholine (5, 10, and 15 μg/min) was infused into the proximal iliac artery to measure endothelial dependent relaxation (EDR). Sodium nitroprusside was infused to determine endothelial independent relaxation at the end of the study to insure smooth muscle vasomotor integrity. External iliac artery luminal diameter was measured using motion-mode ultrasonography. Statistical analysis was performed using repeated-measures analysis of variance with Tukey's post-hoc analysis. Results: All pigs survived the experiment. Pigs required ninefold more resuscitation with LR (370.58 ± 29 mL/kg) versus SB (41.45 ± 3.5 mL/kg) or LRSB (76.4 ± 1.1 mL/kg) (p < 0.05). EDR for LR pigs 1 hour after initiation of resuscitation (R1) was 70.4 ± 14.4% compared with 94.2 ± 13.4% for SB and 106.1 ± 8.2% for LRSB (p < 0.05). At 4 hours after resuscitation (R4), systolic luminal diameters were larger in the SB (0.45 ± 0.01 cm) and LRSB (0.51 ± 0.02 cm) groups compared with LR (0.41 ± 0.03 cm) (LRSB versus LR; p = 0.01). At R4, EDR for the LR group was 78.3 ± 10.7% compared with SB (101.4 ± 8.3%) and LRSB (106.4 ± 7.4%) (p < 0.05). Infusion of sodium nitroprusside confirmed integrity of smooth muscle vasorelaxation. Analysis of serum nitric oxide levels revealed decreased values after resuscitation with LR (9.44 ± 0.76 mol/L) compared with SB (26.3 ± 7.8 mol/L) and LRSB (16.3 ± 1.0 mol/L) (p = not significant). Conclusion: This is the first description of a large-animal model to evaluate EDR after hemorrhagic shock. Resuscitation with LR requires significantly larger volumes than SB or LRSB. LR resuscitation leads to endothelial dysfunction, as determined by decreased EDR, versus SB or LRSB. Resuscitation with blood products may preserve nitric oxide bioactivity when compared with crystalloid resuscitation in the setting of hemorrhagic shock. Copyright © 2005 by Lippincott Williams & Wilkins, Inc.