Rationale and HypothesisAcute lung injury is a frequent complication of severe sepsis or blood loss and is often associated with an excessive inflammatory response requiring mechanical ventilation. We tested the hypothesis that the types of fluids used during early resuscitation have an important effect on the evolution of lung injury.MethodsRats were subjected to either hemorrhage or endotoxemia for 1 hr, followed by resuscitation to a controlled mean blood pressure with Ringer’s lactate, 5% albumin, or 25% albumin for 1 hr. After resuscitation, blood cytokine levels were measured. The lung was then excised and ventilated with a tidal volume of 30 mL/kg for 2 hrs.ResultsThe volume of fluids required was significantly smaller in the albumin-treated groups than in the Ringer’s lactate groups. In the hemorrhagic shock model, plasma concentrations of tumor necrosis factor-&agr;, interleukin-6, and macrophage inflammatory protein-2 were significantly lower and interleukin-10 was significantly higher in the albumin-treated groups compared with the Ringer’s lactate–treated group. The levels of tumor necrosis factor-&agr; and macrophage inflammatory protein-2 in bronchoalveolar lavage fluid were lower and interleukin-10 was higher in the albumin-treated groups than in the Ringer’s lactate group. The decreased cytokine production was associated with a reduction of hydrogen peroxide formation with albumin resuscitation. The lung wet/dry ratio was lower in the 5% albumin (0.54 ± 0.01) and 25% albumin (0.55 ± 0.02) groups than in the Ringer’s lactate group (0.62 ± 0.02; bothp< .05). These effects of albumin seen in the hemorrhagic shock model were not observed in the endotoxic shock model.ConclusionsWe conclude that resuscitation with albumin may have utility in reducing ventilator-induced lung injury after hemorrhagic shock, but not after endotoxic shock. These findings suggest that the mechanisms leading to ventilator-induced lung injury after hemorrhage differ from those after endotoxemia.