Previously, our group developed an ovine model of hyperdynamic sepsis associated with acute lung injury. In this study, we sought to modify this sepsis model by the administration of gentamicin to more closely simulate the symptoms observed in human sepsis in the intensive care unit. In a prospective, controlled, randomized laboratory experiment, 18 female sheep were surgically prepared for chronic study. After a tracheotomy had been performed, the sheep were randomized into sham, control, and gentamicin groups (n = 6 each). Sham animals were surgically prepared for the study but were neither injured nor treated. Control and gentamicin animals received 48 breaths of cotton smoke (<40 degrees C) followed by the instillation (via a bronchoscope) of live Pseudomonas aeruginosa (2-5 x 10(11) colony-forming units) bacteria into the lung. All sheep were mechanically ventilated with 100% O2 for the duration of the 24-h experimental period. Gentamicin (2 mg/kg) was administered at 6, 12, and 18 h after injury. The animals were resuscitated with lactated Ringer's solution to maintain filling pressures and hematocrit on a constant level. Cardiopulmonary variables were stable in sham animals, but in the control group, cardiac index increased significantly after 24 h versus baseline (BL, 5.1 +/- 0.4 L.min(-1).m(-2) vs. 24 h, 7.3 +/- 0.7 L.min(-1).m(-2); P < 0.05); this was associated with a significant drop in mean arterial pressure (BL, 95 +/- 3 mmHg vs. 24 h, 65 +/- 4 mmHg, P < 0.05) and systemic vascular resistance index (BL, 1410 +/- 118 dynes s.cm.m vs. 24 h, 598 +/- 101 dynes s.cm.m, P < 0.05). Treatment with gentamicin stabilized cardiac index (BL, 5.0 +/- 0.4 L.min(-1).m(-2) vs. 24 h, 4.7 +/- 0.4 L.min(-1).m(-2)) and attenuated the decrease in mean arterial pressure (BL, 99 +/- 3 mmHg vs. 24 h, 84 +/- 4 mmHg) and systemic vascular resistance index (BL, 1573 +/- 173 dynes s.cm.m vs. 24 h, 1263 +/- 187 dynes s.cm.m). In addition, the fluid requirement in the gentamicin group was significantly lower than in the control group. Pulmonary function remained stable in sham animals, but the PaO2/FiO2 ratio and shunt fraction deteriorated similarly in the control and the gentamicin groups. Because gentamicin improved hemodynamic variables and reduced the fluid requirement in this ovine model, we believe that this modified sepsis model might provide a clinically relevant and useful new approach for future studies focusing on hemodynamic variables and outcome.