Effect of mechanical ventilation strategy on dissemination of intratracheally instilled Escherichia coli in dogs

Objective: To test the effect of different mechanical ventilation strategies on dissemination of intratracheally instilled Escherichia coli in dogs and to determine the extent and distribution of lung damage. Design: Prospective, randomized study. Setting: Experimental animal laboratory. Subjects: Eighteen anesthetized and paralyzed dogs. Interventions: We studied the effect of three ventilatory strategies based on two variables: transpulmonary pressure and positive end-expiratory pressure (PEEP). Group 1 animals (n = 6) were ventilated with a PEEP of 3 cm H₂O and a tidal volume of 15 mL/kg, which generated an end-inspiratory transpulmonary pressure of ≤15 cm H₂O. In group 2 (n = 6), tidal volume was adjusted to generate a transpulmonary pressure of 35 cm H₂O and PEEP was set to 3 cm H₂O. In group 3 (n = 6), tidal volume was also adjusted to yield a transpulmonary pressure of 35 cm H₂O but PEEP was set to 10 cm H₂O. In each group, we instilled ~10⁸ colony-forming units of E. coli into the trachea of the dogs and ventilated them with the chosen tidal volume and PEEP for 6 hrs afterward. Measurements and Main Results: We measured the pressure volume relationship (pressure-volume curve) of the respiratory system before and 6 hrs after bacterial instillation. We obtained blood cultures before and 0.5, 1, 2, 3, 4, 5, and 6 hrs after bacterial instillation. After 6 hrs, the lungs were removed for histologic (histologic score) and gravimetric (wet-to-dry weight ratio, WW/DW) analysis. During the experiment 0, 5, and 1 dogs developed positive blood cultures in groups 1, 2, and 3, respectively. The number of dogs that developed bacteremia in group 2 was significantly greater than in the other two groups (p < .05). In group 1, pressure-volume curves demonstrated a lower inflection point which was greater than the end- inspiratory transpulmonary pressure suggesting that low transpulmonary pressure/low PEEP strategy ventilated aerated regions without expanding atelectatic areas. In group 2, pressure-volume curves demonstrated both a lower inflection point and an upper deflection point which were spanned by the tidal volume, suggesting that high transpulmonary pressure/low PEEP strategy might have caused both overdistention and cyclic closure and reopening. In group 3, pressure-volume curves demonstrated only a upper deflection point which was teas than the maximal alveolar tidal pressure. At the end of the experimental protocol, group 2 manifested the most lung injury as assessed by gravimetric and histologic indices of lung injury. WW/DW of group 2 (13.1 ± 1.0 (SD); p <.05) was greater than groups 1 and 3 (7.5 ± 1.2 and 8.6 = 1.0, respectively). Similarly, the overall weighted histologic injury score for group 2 (1.19 ± 0.26; p < .02) was greater than for groups land 3 (0.82 ± 0.20 and 0.88 ± 0.22 respectively). For groups 2 and 3, the overall weighted histologic injury scores of the dependent regions were greater than the nondependent regions (p < .004). Conclusions: We conclude that the ventilatory strategy most likely to overdistend the lungs while allowing repetitive opening and closure of alveoli (group 2) facilitated bacterial translocation from the alveoli to the bloodstream and increased lung injury, as determined by histologic and gravimetric analysis. PEEP ameliorated these effects, despite lung overdistention, but increased histologic and gravimetric indices of lung injury in dependent as compared with the nondependent regions.