Prone positioning attenuates and redistributes ventilator-induced lung injury in dogs

Background: We previously demonstrated a markedly dependent distribution of ventilator-induced lung injury in oleic acid-injured supine animals ventil ated with large tidal volumes and positive end-expiratory pressure greater than or equal to 10 cm H2O. Because pleural pressure distributes more unifo rmly in the prone position, we hypothesized that the extent of injury induc ed by purely mechanical forces applied to the lungs of normal animals might improve and that the distribution of injury might be altered with prone po sitioning. Objective: To compare the extent and distribution of histologic changes and edema resulting from identical patterns of high end-inspiratory/low end-ex piratory airway pressures in both supine and prone normal dogs. Design/Setting: We ventilated 10 normal dogs (5 prone, 5 supine) for 6 hrs with identical ventilatory patterns (a tidal volume that generated a peak t ranspulmonary pressure of 35 cm H2O when implemented in the supine position before randomization, positive end-expiratory pressure = 3 cm H2O), Ventil ator-induced lung injury was assessed by gravimetric analysis and histologi c grading. Measurements and Main Results: Wet weight/dry weight ratios (WW/DW) and his tologic scores were greater in the supine than the prone group (8.8 +/- 2.8 vs. 6.1 +/- 0.7; p = .01 and 1.4 +/- 0.3 vs. 1 +/- 0.3; p = .037, respecti vely). In the supine group, WW/DW and histologic scores were significantly greater in dependent than nondependent regions (9.4 +/- 1.9 vs. 6.7 +/- 0.9 ; p = .01 and 2.0 +/- 0.4 vs. 0.9 +/- 0.4; p = .043, respectively). In the prone group, WW/DW also was greater in dependent regions (6.7 +/- 1.1 vs. 5 .8 +/- 0.5; p = .054), but no significant differences were found in histolo gic scores between dependent and nondependent regions (p = .42). Conclusion: In this model of lung injury induced solely by mechanical force s, the prone position resulted in a less severe and more homogeneous distri bution of ventilator-induced lung injury. These results parallel those prev iously obtained in oleic acid-preinjured animals ventilated with higher pos itive end-expiratory pressure.