Augmented lung injury due to interaction between hyperoxia and mechanical ventilation.

OBJECTIVE

Mechanical overdistension and hyperoxia can independently cause lung injury, yet little is known about their combined effects. We hypothesized that hyperoxia exacerbates lung injury caused by large tidal volume ventilation.

DESIGN

Experimental study.

SETTING

University laboratory.

SUBJECTS

Anesthetized, paralyzed rabbits.

INTERVENTIONS

In experiment 1, 12 rabbits were ventilated with 25 mL/kg tidal volumes at positive end-expiratory pressure of 0 cm H2O for 4 hrs with either hyperoxia (HO; FiO2 = 0.5) or normoxia (NO; FiO2 = 0.21). In experiment 2, a separate group of animals were randomized to one of four groups to assess the interaction of tidal volume and inspired oxygen concentration on potential mediators of injury after 2 hrs of ventilation, before significant injury occurs: a) NO+normal tidal volume (NV; VT = 10 mL/kg); b) HO+NV; c) NO+high tidal volume (HV; VT = 25 mL/kg); d) HO+HV (n = 3 per group).

MEASUREMENTS AND MAIN RESULTS

: In the first study, HO compared with the NO group had significantly reduced PaO2/FiO2 ratio (320 +/- 110 vs. 498 +/- 98, p = .014) and increased lung injury scores at 4 hrs. Hyperoxia also significantly increased polymorphonuclear leukocytes, growth-related oncogene-alpha (2073 +/- 535 vs. 463 +/- 236 pg/mL, p = .02), and monocyte chemotactic protein-1 (7517 +/- 1612 vs. 2983 +/- 1289 pg/mL, p = .05) concentrations in bronchoalveolar lavage fluid. The second study showed increased alveolar-capillary permeability to a 70-kD fluorescent-labeled dextran only in response to the combination of both HO and HV. Chemokines and bronchoalveolar lavage fluid neutrophils were elevated in both HV groups; however, hyperoxia did not further increase chemokine or neutrophil counts over normoxia. No difference in lipid peroxidation was seen between groups.

CONCLUSIONS

Moderate hyperoxia exacerbates lung injury in a large tidal volume model of ventilator-induced lung injury. The mechanism by which this occurs is not mediated by increased production of CXC chemokines or lipid peroxidation.