Objectives: To compare four widely used animal models of acute lung injury and to determine the changes in physiologic variables associated with each model.
Design: A prospective, controlled animal study.
Setting: An animal laboratory of a university-affiliated children's hospital.
Subjects: Four groups of anesthetized, paralyzed, and ventilated young Yorkshire pigs, weighing 35 to 45 kg.
Interventions: Acute lung injury was generated by four different methods: a) intrapulmonary arterial infusion of endotoxin of Escherichia colt; b) bronchoalveolar instillation of 0.05N of hydrochloric acid; c) repeated bronchoalveolar warm saline lavage; and d) intrapulmonary arterial infusion of oleic acid. After each acute lung injury procedure, the temporal changes in various physiologic variables were measured, starting at 60 mins and at 15-min intervals thereafter for a total of 165 mins. Systemic and mixed venous serum immunoreactive tumor necrosis factor (TNF)-alpha concentrations were also measured at the same time points. Analysis of variance for repeated measures was employed to determine the absolute and relative significance of the changes observed.
Measurements and main results: Systemic and mixed venous immunoreactive TNF-alpha did not change following any of the acute lung injury procedures. The animals' heart rates and systemic vascular resistances also did not change. Hydrochloric acid instillation as well as bronchoalveolar lavage resulted in significant hypoxemia with no other hemodynamic effects. Endotoxin infusion did not result in hypoxemia but caused significant increases in mean pulmonary arterial pressure and pulmonary vascular resistance and decreases in mean arterial pressure and cardiac output. Oleic acid infusion resulted in a marked hypoxemia with a pronounced increase in mean pulmonary arterial pressure and pulmonary vascular resistance. It also markedly reduced the mean arterial pressure, cardiac output, and the mixed venous PO2.
Conclusions: The surfactant depletion and hydrochloric acid instillation models produce acute hypoxemia in an otherwise hemodynamically stable animal. A brief endotoxin infusion provides a model for cardiovascular instability and pulmonary hypertension but fails to produce hypoxemia in the pig. The oleic acid infusion creates a model of marked cardiovascular instability, pulmonary hypertension, and profound hypoxemia. However, none of the acute lung injury models described was associated with the production of tumor necrosis factor.