Background: Acute respiratory distress syndrome (ARDS) can have various causes. The study objective was to investigate whether different pathophysiologic models of ARDS would show different respiratory, cardiovascular and inflammatory outcomes. Methods: We performed a prospective, randomized study in 27 ventilated ewes inducing ARDS using three different techniques to mimic the pulmonary causes of ARDS (ARDSp): warm saline lavage (n=6), intratracheal hydrochloric acid (HCl; n=6), intratracheal albumin (n=10), and one technique to mimic an extrapulmonary cause of ARDS (ARDSexp): intravenous lipopolysaccharide (LPS iv; n=5). ARDS was defined when PaO 2 was <15kPa (112 mmHg) when ventilated with PEEP 10cm H 2 O and FiO 2 =1.0. The effects on gas exchange were investigated by calculating the oxygenation index (OI) and the ventilation efficacy index (VEI) every 30 minutes for a period of 4 hours. Post mortem lung lavage was performed to obtain broncho-alveolar lavage fluid (BALF) to assess lung injury and inflammation. Lung injury and inflammation were assessed by measuring the total number and differentiation of leukocytes, the concentration of protein and disaturated phospholipids, and interleukine-6 and -8 in the BALF. Histology of the lung was evaluated by measuring the mean alveolar size, alveolar wall thickness and the lung injury score system by Matute-Bello et al., as markers of lung injury. The concentration of interleukin-6 was determined in plasma, as a marker of systematic inflammation. Results: The OI and VEI were most affected in the LPS iv group and thereafter the HCl group, after meeting the ARDS criteria. Diastolic blood pressure was lowest in the LPS iv group. There were no significant differences found in the total number and differentiation of leukocytes, the concentration of protein and disaturated phospholipids, or interleukin-8 in the BALF, histology of the lung and the lung injury score. IL-6 in BALF and plasma was highest in the LPS iv group, but no significant differences were found between the other groups. It took a significantly longer period of time to meet the ARDS criteria in the LPS iv group. Conclusions: The LPS model caused the most severe pulmonary and cardiovascular insufficiency. Surprisingly, there were limited significant differences in lung injury and inflammatory markers, despite the different pathophysiological models, when the clinical definition of ARDS was applied.