PO(2) oscillations induce lung injury and inflammation

BACKGROUND: Mechanical ventilation can lead to ventilator-induced lung injury (VILI). In addition to the well-known mechanical forces of volutrauma, barotrauma, and atelectrauma, non-mechanical mechanisms have recently been discussed as contributing to the pathogenesis of VILI. One such mechanism is oscillations in partial pressure of oxygen (PO(2)) which originate in lung tissue in the presence of within-breath recruitment and derecruitment of alveoli. The purpose of this study was to investigate this mechanism’s possible independent effects on lung tissue and inflammation in a porcine model. METHODS: To separately study the impact of PO(2) oscillations on the lungs, an in vivo model was set up that allowed for generating mixed-venous PO(2) oscillations by the use of veno-venous extracorporeal membrane oxygenation (vvECMO) in a state of minimal mechanical stress. While applying the identical minimal-invasive ventilator settings, 16 healthy female piglets (weight 50 ± 4 kg) were either exposed for 6 h to a constant mixed-venous hemoglobin saturation (S(mv)O(2)) of 65% (which equals a P(mv)O(2) of 41 Torr) (control group), or an oscillating S(mv)O(2) (intervention group) of 40–90% (which equals P(mv)O(2) oscillations of 30–68 Torr)—while systemic normoxia in both groups was maintained. The primary endpoint of histologic lung damage was assessed by ex vivo histologic lung injury scoring (LIS), the secondary endpoint of pulmonary inflammation by qRT-PCR of lung tissue. Cytokine concentration of plasma was carried out by ELISA. A bioinformatic microarray analysis of lung samples was performed to generate hypotheses about underlying pathomechanisms. RESULTS: The LIS showed significantly more severe damage of lung tissue after exposure to PO(2) oscillations compared to controls (0.53 [0.51; 0.58] vs. 0.27 [0.23; 0.28]; P = 0.0025). Likewise, a higher expression of TNF-α (P = 0.0127), IL-1β (P = 0.0013), IL-6 (P = 0.0007), and iNOS (P = 0.0013) in lung tissue was determined after exposure to PO(2) oscillations. Cytokines in plasma showed a similar trend between the groups, however, without significant differences. Results of the microarray analysis suggest that inflammatory (IL-6) and oxidative stress (NO/ROS) signaling pathways are involved in the pathology linked to PO(2) oscillations. CONCLUSIONS: Artificial mixed-venous PO(2) oscillations induced lung damage and pulmonary inflammation in healthy animals during lung protective ventilation. These findings suggest that PO(2) oscillations represent an independent mechanism of VILI. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s13054-019-2401-1) contains supplementary material, which is available to authorized users.