Experimentally, the intravenous administration of a bolus dose of Escherichia coli endotoxin in sheep or a bolus dose of live Pseudomonas aeruginosa in rats is insufficient to cause injury to the alveolar epithelial barrier. Therefore, the first objective of these studies was to maximize the injury caused by live bacteria to the lung by administering a large dose of live P. aeruginosa into the lung perfusate of goat lungs in situ. P. aeruginosa (2.4 x 10(10) colony-forming units [cfu]) and [131I]albumin (vascular protein tracer) were added to the lung perfusate. Even though the bacterial inoculum remained very high in this isolated perfused lung system, there was no change in the permeability to protein or clearance of fluid across the alveolar epithelium, although there was an increase in lung endothelial protein permeability. Therefore, since systemic factors have been implicated in the severity and pathogenesis of septic lung injury, the second objective was to administer a continuous intravenous infusion of live P. aeruginosa over 8 h in intact anesthetized sheep. The eight sheep so treated exhibited an intact, functional alveolar barrier, even though there was an increase in lung endothelial permeability to protein and an increase in extravascular lung water. In fact, in these eight sheep, alveolar epithelial fluid transport was significantly greater than in control sheep. In the three other septic sheep there was injury to the alveolar epithelial barrier with an increase in permeability of the barrier to protein, an inability to transport fluid out of the airspaces, and an even greater increase in extravascular lung water.(ABSTRACT TRUNCATED AT 250 WORDS)